As a plugged-in reader notes, Fortune Magazine looks at historic Price to Rent ratios (pages 76-88) across 54 major metropolitan areas in an attempt to forecast where home prices are headed.
In most markets people won’t lay out much more in monthly costs to own a house or condo than they would to rent a similar property unless they expect a huge profit when they sell. Indeed, speculators chasing quick profits did a lot to inflate the recent bubble. But once the fervor fades, prices must fall to restore their normal, long-term relationship with rents. Rents exercise a kind of inevitable gravitational pull on prices. The ratio of prices to rents “behave much like price/earning rations for stocks,” says Yale economist Robert Shiller. “Like P/Es, price-to-rent ratios are mean-reverting.” In other words, while prices soar from time to time, sending the ration to exceptional heights, sooner or later the relationship is bound to return to its historical average.
Of course, rather than prices falling rents can rise. And while Fortune forecasts a 28.3% correction in the Price to Rent ratio over the next five years for the San Francisco MSA, they’re forecasting it’s derived from a 10% drop in prices along with an 18% increase in rents. Over in the East Bay, however, the Fortune forecast is a bit more grim as they’re calling for a 38.1% correction driven mainly by price declines of >30%.
And of course, the concept of a housing P/E (and rising rents) shouldn’t catch any plugged-in readers by surprise.
Over in the East Bay, however, the Fortune forecast is a bit more grim as they’re calling for a 38.1% correction with over 30% coming from price declines alone.
Should be: “30 percentage points coming from price declines alone.”
[Editor’s Note: Unclear wording revised (and thank you for plugging in).]
See, I told ya Fluj!
Link to a relevant article:
http://globaleconomicanalysis.blogspot.com/2006/08/beware-of-alligators.html
Interesting stuff. You cut off the far left-hand columns. Forecast for SF “Value of an upscale home (one that sells for double the local median price) in five years” —
2007: $1,730,000; 2012: $1,568,000.
Fortune’s economists actually have SF housing holding it values better than the CSI futures markets at the Chicago Merc — traders are putting the decline in SF at about 25% over the next 4 years. You can make a lot of money if you think values will hold better than that (if you’re willing to put your money where your mouth is).
[Editor’s Note: With regard to San Francisco and the CME futures market: They’re Betting Against Us (San Francisco) On The CME.]
How does rent control affect these numbers? Won’t it take like 12 years for rent-controlled SF to increase rents by 18%?
Presumably, they are comparing the price of buying a home to the rent on a newly-rented home. While rent control stabilizes rent for those who remain in their units (built before 1979), there is no vacancy control, so rents in vacant homes rise to what the market will bear.
I told ya so. LOL. This isn’t all that dire of a forecast for SF if you look closely. I’ve said all along. that some sort of correction is likely. But come on. Fortune is no different than anybody else. It’s always coming sometime soon, everybody thinks. Nobody knows when.
I wonder how they came up with their estimate for rising rents. I didn’t see the methodology in the article. 18% seems low over a five year period. Rents were up around 12% last year alone.
“18% seems low over a five year period. Rents were up around 12% last year alone.”
Ditto that anono — just don’t see how anyone can think it’s going to take 5 years to work this alleged bubble out. Seems just as likely that the market could bottom out Q3 of 08, rebound for a couple of years and head south again by then. No one has that kind of crystal ball with the number of variables in play.
I’m moving to Cleveland!
Rents cannot rise above a level supported by wages.
You can get a short term spike, but because of renters greater mobility they will move out of the area rather then pay the increased rent and rents will return to a level supported by incomes.
This is why monthly rent payments are part of the CPI and mortgage payments are not.
Unless wages rise 18% rents cannot rise that far that fast.
And if wages rise 18% that is a big chunk of the necessary correction to bring home prices back into line with historic levels of affordability.
OK, then. Does anyone care to predict by what month the median price, year over year, will start to fall in SF? In 10 of the past 11 months, the SF median price has been higher than the same month in 2006. When is the alleged downturn gonna start, folks??
Don’t worry sanfrantim, the drop is coming. It’s already here in some crappy outer neighborhoods. Why are you so anxious for prices to drop, anyway??
Prices down, rents up– sounds like no one’s gonna be happy in that scenario….
Not anxious, Jimmy. Skeptical.
I thought there was no more buildable land in the Bay Area- even in the East Bay? That’s been one of the arguments the past 7 or 8 years for the sustainability of the housing boom. It’s not as if there can be massive new housing developments in Berkeley anytime soon. Why is it that housing prices are going to fall there by such a large percentage?
It doesn’t make any sense if you follow the old argument about lack of land.
Rents cannot rise above a level supported by wages.
For a region, sure. But there’s nothing to say that socioeconomic shifts within a region cannot happen, causing rents in one area to greatly EXCEED growth in incomes of the region. For example, if more DINKs, singles, and retirees consider renting in SF to be worth more than renting in SJ – rents could rise more here as more affluent peeps push out the less affluent. Salaries wouldn’t have to change at all to significantly push rents in one small part of the larger metro significantly higher.
Not saying it will happen – just saying that it’s very simplistic to apply macro views to a micro area.
It’s not as if there can be massive new housing developments in Berkeley anytime soon. Why is it that housing prices are going to fall there by such a large percentage?
Read the article. The article regularly refers to the “East Bay” as Walnut Creek. They aren’t clear exactly which areas are covered by “East Bay”. Don’t expect prices in Berkeley to drop by anywhere near the amount new McMansions in Wally Creek will.
“Lack of land” doesn’t really mean anything here in SF. New condo buildings are rising higher and higher. Permits to exceed previous height limitations have been handed out left and right. As to “when is the alleged downturn” going to start, I don’t know but perhaps when escrow begins closing on all those projects downtown and the buyers realize they can’t get the terms they thought they were going to be able to get back a half a year ago, if they can get a mortgage at all?
Who cares about the median price…it doesn’t mean anything anyway as it is skewed by a smattering of low priced and extremely high priced sales.
dollivm —
actually, I think you are referring to the mean price which is an average and could be affected by a few high priced sales outside the norm.
median price is used to correct this — it takes the middle price in a range of numbers.
that’s why you’ll see median more than you will mean or average, because it’s a more accurate reflection of prices.
Median is clearly better than average, but given the divergence of medians and the Case Shiller index in most areas that that index covers, it’s pretty obvious that median isn’t necessarily a great indicator of what’s happening right now in those areas:- presumably too much change of “mix” in sales due to the subprime mess.
And, no, I’m not arguing that SF prices are declining, and yes I’m aware that Case-Shiller for SF doesn’t just cover SF 🙂
Anyone who predicts what prices, rents and population will be in one to five years is just speculating. Are you going to tell me that someone who is able to predict how many jobs will be in the Bay Area, what the average income is going to be, the average rents, the market cap of the companies in the bay area are, the population, the Dow, the Nasdaq, and the US Dollar is going to be explaining it to us by writing an article in Fortune? That’s just ridiculous. The only certainty I know of: Population goes up + Income of Popultation goes up = higher prices.
I’ve been waiting too for some clear, unequivocal year-over-year price declines on a per unit (or per sq ft) basis. I don’t think we’ll get it with the Oct 07 v Oct 06 data, but I do think that we’ll see actual declines soon. The continuing rise in rents is also surprising – at least the double-digit magnitude. As many have pointed out here, markets are made at the margin. And given our constrained supply, there seems to have been enough purchasing power at the margin to keep prices high. This may be played out though – particularly given new tighter lending standards and at least some new supply coming on line. Reversion to the mean is a pretty powerful concept.
Re: markets made at the margin. Rents for vacant apts can go up 18% without wages going up 18%, because most units are occupied, with rent-controlled, below-market rents.
When people with higher incomes are moving to the City, as they did during the dot com boom and as they are now, all renters don’t need to afford the higher rent– just the small percentage that are new renters.
Of course, this price volatility works the other way, too. During the dot com bust, rents quickly dropped again.
The maginitude of the estimated rent increase for SF is huge. Nobody wins here, but so far, rents are up pretty big this year, and prices haven’t declined in SF yet at all from what I’m observing.
“Fortune’s economists actually have SF housing holding it values better than the CSI futures markets at the Chicago Merc — traders are putting the decline in SF at about 25% over the next 4 years.”
The Fortune prediction is not that different from the Chicago trading of CSI futures, since that is a prediction of the price of single family homes throughout the Bay Area, including the East Bay. The number of single family home sales sold in SF is a small fraction of those sold in the East Bay, so the Chicago trading of CSI futures is closer to the Fortune prediction for the East Bay than for SF proper.
The numbers for SF look dead-on to me, if not the exact time frame.
Anecdotal evidence suggests something like 20% rent increases over the past year. A 1br went from 1400 to 1700, for example.
A 10% drop in house prices would bring the P/R from 38.2 to (0.9 / 1.2) * 38.2 = 28.65. Sounds pretty plausible.
10% is still a significant drop for SF where many homes are highly leveraged.
MORE dinks and singles??? is that possible???
No surprises here. We left the EBay for Fort Lauderdale because housing costs were so extraordinary. FWIW…Fort Lauderdale, with a population 1/5th San Francisco, has 10 times the number of MLS listings. It’s a buyers nirvanna here.
Y/Y prices have fallen in SF, if you are one of the increasing numbers unable to sell your home at an “acceptable” price (and thus withdrawn it from the marketplace).
Cary
You don’t need to go all the way to FL to find lots of listings. Sacramento County has about 10X the number of listings of SF with a population similar to Broward County FL. Parts of East Bay too. So buyers nirvannas are getting pretty close to Fortress San Francisco.
Interesting that East Bay, San Jose, and S.F. have the highest current multiples of the 54 metro areas, followed closely by Honolulu. Similar pattern with the 15-year average multiples.
The 15-year averages are skewed upwards by the most recent several years. If one believes in mean reversion (and the possibility of overshoot), the multiple for SF could fall well below 27.4.
BTW, if home prices are falling, it doesn’t mean that rents will rise. I recall in the early 90s, people moved away from the Bay Area due to the perception that housing was too expensive(!). Can’t find rent growth #’s for the early 90s but I note that average rents in the City grew only 1.1% and 1.7% during 1993 and 1994 (source: RealData).
In the event of a recession, rents could stagnate or decline and multiple compression could have a more exaggerated effect on home prices.
That’s not a prediction, just a possibility.
This article is full of statistical fallacies:
1. The choice of a 15 year average is entirely arbitrary. Why not a 5 or a 50 year average?
2. Any data series can be shown to be mean reverting but that does not actually mean that it must or will revert to a mean. For every year that it doesn’t revert, the mean of the now expanded time series will change. Prices could rise more slowly than they did in the historical sample and the mean would climb faster than the YoY change in prices. Voila! Prices revert (upward) to the mean.
3. These houses are in not apples to apples. If the footprint, finishes, energy efficiency etc. of newer housing stock is superior to the houses of 15 years ago, then all else being equal, it would be worth more for the fundamental reason that it’s better quality housing.
4. Other moving parts include changes in tax policies, crime rates, yields on alternative investments, etc., etc., etc.
Waffle: I admire Schiller’s work.
Unwaffle: This isn’t one of his better insights.
I rented for years until I was threatened with the sale of my rental and realized I had thrown away almost $200,000 and nothing to show for it. Now prices are falling and I’m probably losing at least that again! I don’t consider myself dumb and I don’t believe I could have done things any differently, sometimes things are not within your control. For all of the those predictions of home prices coming down 25%, that is also tied to some hard economic times overall. I thought the comment questioning whether it was possible to increase the number of dinks and singles pretty funny.
Viewlover, it’s not necessarily true that just because you paid almost $200,000 in rent over a number of years you “have nothing to show for it.” With the money you did not have to put down on a purchase and the money saved from the much lower monthly costs to rent rather than own in recent years, you could have had very nice returns in the markets. You very well may be ahead if you do the analysis.
Vox, I have to respond to your points:
1. Yes, a 15 year average is relatively arbitrary, but as SFAnalyst pointed out, this benefits the pricing imbalance calculation due to the past 8 years being completely out of whack from the past 100 year history. If he chose a 50 year average, the predicted adjustment would be much worse.
2. Don’t really know what you’re saying here, but Schiller agrees that you can revert to the mean P/R by either having rent increases or real estate declines or a combination of both. So yes, price stagnation without decline will likely help lower the P/R ratio over time.
3. This is for the most part not true. You can surely increase the value of an individual house by upgrading, but you cannot increase the value of a market through upgrades. As the report shows, there is a fundamantal and relatively unchanging balance between income and rent prices over time (despite the past 7 years’ wackiness), as well as rent prices and housing prices. If every single house in San Francisco had $50,000 in upgrades done in the next month, the value of each house would not go up by $50k. There would still be better houses and worse houses, and the worse houses would have wasted their money on upgrades because they’d still be stuck having to price at the low end of the market. Every seller still needs a paying customer, and if every seller turns their houses into $1.5 million beauties, guess what, the price of $1.5 million beauties will go down because there are too many of them.
So yes, you can buy an entry level house and upgrade it significantly to boost it to a luxury home, but someone always has to be entry level, so you better hope all the joneses on your block don’t get the same idea at the same time that you do.
that’s a huge assumption trip. i’m sure he spent it all fast cars and loose women, as the saying goes. not that there’s anything wrong with that.
That’s why I qualified everything with “could” and “may.” But if he did spend it all on fast cars and loose women, then his conclusion that he has nothing to show for it is even further off the mark . . .
I’m a bear overall on the SF market, but having witnessed the NYC market grow beyond all comprehendable statistics; I think that all of this market prediction is pure speculation with some hard data points to sway the reader. FWIW, I’m still renting and don’t plan on changing that position despite have the means to do so if I chose. I just don’t think it’s a wise use of funds.
@vox
As you point out, there is some art to statistics. How do you pick a representative time period? My guess is that 15 years is more representative than 5. After a certain amount of time, consumers adjust to price changes so a 50-year average may also be unrealistic.
I agree that means can shift permanently due to exogenous changes (e.g. technology, demographics, prices of other goods/commodities, growth of industry, rise/fall of national economies, etc.). The multiple expansion of the past 5 years seems to be a result of a temporary change in how homes were financed. As the financing of homes is now reverting to more traditional standards, it seems likely that prices would follow (barring some external factor).
Even if changing homebuyer preferences showed a willingness to devote a higher percentage of income to housing, that would be irrelevant if bankers were unwilling to lend on that basis.
Finally, wrt your comment about new homes being “better” than old homes. A possibly true (debatable) point. As other commenters have pointed out, we need a housing stock that people can afford. Without an ever-increasing supply of rich people, where are the buyers for these better homes?
A relative lives in a 1,800-SF tract home in Tiburon purchased 10 years ago in the low-$400s. Probably “worth” $1.2MM at the peak in mid-2006. Their family could not afford to live there if they had to pay that price. The same may be true of many of their neighbors.
The most obvious factor in P/R ratio is the mortgage rate. The 15-year average of the 30-year mortgage is about 7.5%, about 1% higher than the current rate. That factor alone will count for about 8% in P/R ratio (across the board, in CA and in the whole country).
I have pointed out in another thread – tax deduction is another factor. When people’s income move up, they get into high marginal tax bracket and the tax deduction becomes more significant.
And yes, finally there is a factor of new contructions and remodeling. Rental properties are often neglected, and if any repair is done, it is the cheapest kind. Owner occupied properties are often remodeled with high quality material and labor. Just take a look at the listings and see how many of them are recently remodeled.
Nobody would remodel a condo/house so he can rent it out. However, most people will do at least some kind of remodeling to make it more appealing when they sell.
The key factor is the “P/R ratio of the particular property”, instead of “Median price”/”Average rent”, because they are apple and oranges, especially in SF.
When will the correction start? Prices have already fallen in many areas of the city and continue to trickle lower. I’ve seen a steady increase in foreclosures/short sales in most neighborhoods, but obviously more pronounced in the outer areas. These will be your downside comps because, of course, the greedy (or overlevered) sellers “aren’t just going to give it away.” But the banks will.
Anyone see the article in the Marin Journal a few days ago? 14% of listed inventory in Novato is in some form of distress.
http://www.marinij.com/marin/ci_7365731
This article makes a great point near the end – the areas hardest hit are the ones where people actually pay their mortgages by working real jobs. Is there enough wealth in the city to prevent price drops in certain areas? Maybe. Luckily for me, I’m not looking to buy in those areas.
Eddy,
NYC has 8.2 million people living in a 362 sq. mile area. SF is 47 sq. miles and had no where near the population density if NYC.
SF pop. Density. 15,834/ sq. mile
NY 27,083/sq mi
Unlike NY where manhattan has a concentration of the extremely high paying jobs in the NY metro area. Most of the Tech Jobs are not in SF proper but further south. NY also has a far superior and cheaper mass transit system.
Given all these factors I don’t think you can compare NYC to SF directly.
Man, if rents really went up by 18% (and prices fell), a lot of condo owners would rent out their existing places for a profit and buy another unit on the cheap as a principal residence. Sounds like a great opportunity to be a landlord. Such high rents would cover my mortgage, HOA, property tax, and car payment. Bring it on!
John,vox,
Hasn’t it always been the case that properties for purchase are of higher quality than rentals? Or has that changed in the last few years?
A house built in the 80s would use newer materials and tech than one from the 60s. so on and so forth. We didn’t take a quantum leap in fittings and material tech used in houses in the las few years did we? New tech makes things cheaper to produce and brings costs down over time.
None of those factors explain the sudden increase in P/R ratios over the 15 year avg..
akros, your analysis might be a bit flawed. check this out:
http://en.wikipedia.org/wiki/Manhattan
i would not compare sf’s 7×7 to anything but manhattan proper. including the boroughs makes it way skewed.
dude,
i did read that online yesterday. very interesting. the challenge for all you sideliners here in the city happens to be that there are plenty of folks with the means that will scoop up anything turn key in a nice hood as soon as it becomes available. these are not working class people, they are well healed.
James,
Ok let’s compare Manhattan shall we. Including the boroughs actually helps SF.
Manhattan Pop, Density: 66,940/sq mi
SF pop. Density. 15,834/ sq. mile
According to the average weekly wages of both places
http://www.bls.gov/ro9/qcewca.htm
New York County, N.Y., held the top position with an average weekly wage of $1,781
San Francisco. $1,481
Santa Clara county. $1,569
It is quite clear Manhattan has a 4.4 times the number of people / sq. mile than SF proper. Not only that people in NY make more on average per week.
In fact, like I said the tech companies are further south as explained in the Santa Clara wages.
So you can’t make any sort of prediction of SF prices based on Manhattan as a reference model.
OK, so here’s a nuts and bolts example of why it’s fallacious to argue mean reversion; and, why it’s silly without disclosing standard deviations; and, why it’s even sillier if there isn’t a discussion characterizing the probability distribution (Normal, leptokurtic, etc.)
In the market for 10 year Treasury bonds, over the last fifteen years, the average yield was 5.39% and the standard deviation was 1.05. A la Schiller, I would say that at a current yield of 4.32% that the bond market is overvalued by 19.8%. Look out below! Not only that, but it’s now trading just over one standard deviation above the 15 year mean making current prices rather improbable.
How about if I use 45 years of data? Then the mean yield is 7.05% and the standard deviation is 6.42%. On that basis, bonds are 38.7% overvalued. But wait! Now they’re also well inside one standard deviation so it’s not at all improbable that they’re this far from the 45 year mean.
How about if I used the data from ’62-92 to predict what would happen next? In that case, the mean was 7.87 and the standard deviation was 7.06. That would’ve told me to look for prices to move down as yields would need to back up from their 6.X% ’92 levels to revert to the mean 7.87%. Wrong! Over the ensuing 15 years bond yields only traded above the 62-92 mean for one month. Meanwhile they went as low as 3.35%. I would’ve mised one whale of a good bond market. Not only that, but the prior high volatility, (esp. as shown in the 70s) fell hugely, as shown by the relatively tiny standard deviation of yields from 92-07. Not only did prices go down, so did the innate riskiness of the bonds.
All of the above is based on the idea that bond yields have a normal distribution. How would I guess that? From the data set. But how would I know if the data set was adequate? From the probability distribution inferred from the prior data set. Uh oh. Circular reasoning.
i wasn’t arguing with you that nyc is much more populous or wealthy, just that you shouldn’t include the boroughs in a purist exercise. we don’t have any boroughs here unless you consider the sunset or bayview to be ours. i’ve often felt we should annex south sf, brisbane and daly city to make them true boroughs and start to finally try and address affordable housing where land is more realistically usable for such projects.
as for nyc, i do think we are headed that way, wherein, you make millions per year or you are just trying to survive like everybody else.
Connecting the dots, a lot of the divergence from the 15 year rental yield is fundamentally accounted for by a comparable divergence from the 15 year mean in the Treasury yields. Unless Schiller wants to argue that the rental divergence is several Sigmas away from the mean (that’s doubtful) then prices are not improbably high.
And here is some fun statistical porn from the Fed (yes, it’s outdated, pre 21st century) showing the changes in the hedonic (qualitative) aspects of the housing stock, amongst other costs of living. Here the relevant mean isn’t an absolute number but hours worked to purchase a sq. ft. As of the late ’90s this hadn’t moved much but the market had brought in larger footprints, HVAC, garages, dryer hookups and other improvements previously unknown to consumers without budging hours worked/sq. ft.
See:http://www.dallasfed.org/fed/annual/1999p/ar97.pdf
Vox, not sure what it is you’re implying. A paradigm shift in the fundamentals of local real estate? The only one I’ve seen is people using loans they shouldn’t have gotten to pay too much for property they couldn’t afford in the first place. And that would indeed skew a distribution.
Rather than looking at treasury yields, try re-running the same analysis on corporate bond spreads, notably high yield, to treasuries and see what the pattern looks like.
Anyway, are you saying the real estate P/E relationship is not mean-reverting? If San Francisco was an lunar colony I would agree with you. But given any of us can move to other cities at any time (the east bay is a 25 minute BART ride), it’s unlikely that the rest of the bay area sees an adjustment and we don’t.
good post vox. I agree that it is not helpful to use 15 years of data when everyone agrees that the last 7 of those years have been very out of the ordinary- that’s half the sample. However, I’m sure the counter argument is that p/r ratios are a self contained analysis, whereas bond prices are influenced by so many factors that it isn’t accurate to describe them as having a normal distribution in the first place.
James,
I am actually glad you brought up manhattan. It shakes up most peoples notions that SF is some how a Manhattan clone and thus deserves to have the same ridiculous cost of living.
The conditions around SF are very different. Most people have to commute to Manhattan for work. In the Bay Area that work is spread out in a much larger area. It would be hard for many people to have to deal with traffic every day commuting to work in Santa Clara and living in SF. The public transit system is horrendously inefficient and expensive in the Bay Area. In fact, I have a few friends that have moved out of SF for that very reason. Eventually convenience and kids win over the desire to live in a City.
If SF becomes too expensive most people will just move closer to work and live in more luxury. Also companies would then as well setup shop in the suburbs to save cost.
Anono makes a good point: “it is not helpful to use 15 years of data when everyone agrees that the last 7 of those years have been very out of the ordinary- that’s half the sample.”
So Fortune is being conservative. To revert to the “true” mean, we will need to see a larger correction.
akrosdabay,
While the owner occupied units have always been of better quality than rentals, the difference has become more significant. Anyone who has been tracking the market and going to the open houses would know. Two years ago, the houses are cleaned up and repaired, but I didn’t see that many completely remodeled. Now, it feels a lot (1/4? 1/3?) had significant remodels.
As the difference of quality changes between the “for sale” and “for rent”, the “median price/average rent” ratio will go higher. However, if you really look at the “P/R ratio of the same unit”, it doesn’t move that much.
@Dude, I prefer John Donne, but hey, I think “lunar base” really does get the point across that not everyone is Ess Eff provincial. I mean, I am Berkeley provincial (uh, why would you want to live anywhere else), but I realize that most of the folks I know DID make the decision to live one burgh over (mostly because of housing costs).
Here are some additional stats to chew on from the HSBC report A Froth Finding Mission: Detecting US housing bubbles. You can find it at http://www.capitalism2.org .
Price to Income Ratio
Q3 2005 20 yr avg 30 yr avg
Oakland MSAD 11.6 6.6 5.9
SF MSAD 11.3 7.2 6.4
Price to Rent Ratio
Oakland MSAD 41.6 25.5 23.8
SF MSAD 32.7 21.3 19.7
Actually I did spend it on fast cars and the tech bubble. My BMW 840 was a big lemon and the tech market bombed in 2000. My new condo costs about twice as much as my $2,000 rent after taxes, with 30% down, but it is quite nice, new construction in Corona Heights and 1,800 square feet with very nice finishes. Can’t complain about the lifestyle, just that it is expensive and you are not guaranteed any return on “investments”, plus it can cost more than you bargain for.
If you trend treasuries to themselves then you won’t get useful information. That is like trending housing prices to themselves.
If you trend treasuries to corporate notes you will notice that they always revert to a historic ratio as the market recalculates risk after coming off a risk averse bubble high (or going into one).
The Fortune article makes fundamental sense because as so many people have pointed out – the economics of living only allow people to afford so much for shelter period. The rent for say a 3 bedroom may be higher in SF than in Dallas due to constraints such as limited availability and demographic data (young professionals open to roommate scenarios) but that has little bearing on the historic norm of the ratio.
Taking 15 years worth of data makes sense because of deviations causing short term trends.
John,
The “flipper” culture has only been prevalent in the past few years because the notion of perceived “wealth” has been prevalent.
The full extent of the decadent ways fueled by all this fake wealth from a speculative housing market is unfurling in wall street and the banking system. Even at the global scale, the asian markets went down on US Subprime concerns. It will make its way to the bottom line of more and more peoples income.
I wouldn’t expect what has happened in the last couple of years to be sustainable permanently. Eventually the point of diminishing returns sets in and no amount of remodeling will sell a house at inflated prices. Because perceptions of the value of the remodel depends on how much “perceived” wealth a person has and how much of it was created by the housing industry over the past few years.
akrosdabay,
The flipper culture and the bubble feed each other. However, all I am saying is that the P/R ratio is only meaningful if you use the same particular unit.
When when the bubble burst, the reverse of flipper culture will drop the P/R ratio. Of course, the bears will say that’s the indication of bottom, but when you look at a particular unit, again, you won’t see that much difference.
Also, the Price/income ratio has the same problem. It is only meaningful if you use the price and the buyer’s income (instead of median price/median income), especially in SF where 70% rent.
Since only 30% of SF households can buy, the median BUYER’s income would be about top 15% of the overall households, so the P/I ratio change may indicate the widening gap between the high-income households and the average households.
I would like to see some data on the P/I ratio of each purchases.
“Flipper” culture has only been known to the masses in the past few years. Tradespeople have been doing it for generations. San Francisco is paradise for carpenters. Nowhere else within thousands of miles are there so many decrepit (in realtor speak “deferred maintenance”) little granny houses, often on terrific blocks.
“I would like to see some data on the P/I ratio of each purchases.”
As I’ve said before, if we had this information, there’d be nothing to debate about here on SocketSite. But given that 2/3 of the mortgages in San Francisco are adjustable or interest only, it’s highly likely that many current owners bought more than they could afford.
@Wallstreeter
“If you trend treasuries to themselves then you won’t get useful information. That is like trending housing prices to themselves.”
Technically, I’m looking at yields, not prices, and doing the exact same mumbo jumbo as Schiller did. Whether one is analyzing one number, or a spread between two numbers, the statistical fallacies of forecasting mean reversion from an historical sample are the same.
“If you trend treasuries to corporate notes you will notice that they always revert to a historic ratio as the market recalculates risk after coming off a risk averse bubble high (or going into one).”
For any historical data set, there will be a high, a low and a mean. They quite certainly do not always revert to an historic ratio. Had we played this game in the equities market using pre ’90s data we would’ve suggested that at a 16-18 P/E the market was at an inviolable ceiling and must revert to 13. Instead it then plotted valuations beyond any formerly observed. LTCM blew itself up by getting wrong several different bets, anticipating the narrowing of what were then historically wide spreads. Oops! They widened some more.
Or we might’ve forecast in the 60s that money market rates couldn’t hit 15% because it never happened before.
Or the Dow couldn’t go down 23% in one day in 1987 because that had never happened before.
Schiller, Forbes, et. al. don’t go as far as to show whether the current P/R is even beyond one Sigma off the mean. Neither do they show that the probability distribution is normal as would be necessary to take their forecast of falling prices and rising rents seriously (perhaps that’s because signs point to a Weibull distribution?).
vox – what are your thoughts on the Eichholtz study?
“But given that 2/3 of the mortgages in San Francisco are adjustable or interest only, it’s highly likely that many current owners bought more than they could afford.”
Unfortunately, I won’t agree with you here. Those are good options even for the people who can easily afford a 15-year fixed.
If I can make more money on the stock market, why not use interest-only? If I play to live there for 5 to 7 year only, why not use 5 to 7 year ARM?
Of course, there are people who bought more than they can afford. However, we don’t know how severe it is. Too many variables here.
I just like to say the opposite of the main media. When the media is praising the health of the RE market two years ago, I was a huge BH. However, now it all turns into negative, I want to talk some reason into this.
Vox, What you are describing is a standard fluctuation which is the entire point of taking 15 years worth of data. If the study only took the last 5 years worth of data or took 5 years of data from 90 to 95 then you would have a point.
If we follow your conclusions then the PE for any given stock will only go UP UP and AWAY!
Shoot.. by that logic then at any PE a stock is a buy because it will keep going up. Also, any price for a home would be a bargain because it doesn’t matter how little people are paying for shelter when they don’t own a place (rent) because there will be infinite demand regardless of price for owning.
“For any historical data set, there will be a high, a low and a mean.”
That’s a good point.
In statistics, we would plot the data and try to figure the mean, and the standard deviation. The mean itself is meaningless. As long as the data is withint one or two standard deviation, it is considered normal.
Also, if you do any stock trading at all, you will know that you cannot forecast future trend by looking at the P/E ratio. Some stocks are bad buys even though the P/E ratio is low. Some stocks are good buys even though the P/E ratio is high.
We are talking a lot of nothing here.
“Some stocks are bad buys even though the P/E ratio is low. Some stocks are good buys even though the P/E ratio is high. We are talking a lot of nothing here.”
Uh, no. As our realtor friends are fond of telling people, you can’t live in a stock. Everybody has to live somewhere, and you have the choice of renting or owning. Since the end of the boom, rents have risen while sales numbers have fallen dramatically and values have started to come down. Fortune predicts this will continue. Why? Because you MUST either buy or rent, and it currently makes more economic sense to rent. That’s the whole point of the discussion, John. But thanks for the entertaining attempt at reductio ad absurdum.
akrosdabay, my point on NYC is that prices can continue to go up despitethe fact that we are at all time high’s here in SF; and NYC proves that rents and costs can increase much further than people believe if the underlying circumstances are correct. Bottom line is that in Manhattan you cannot get anything for under 900/psf except maybe for the very unusual place; and this doesn’t factor in the maintenance on most places that exceeds north of $1k per month. For 900/psf in SF you can get a very nice place and maybe even a little bit of a view; perhaps a garden and a parking spot.
I disagree that NYC is not a comp to SF. The two cities are comparable but you may have to apply a slight discount to SF; perhaps SF’s income per capita is lower, but SF provides the only viable urban lifestyle in NorCal; and the type of individual that would like to live in an urban environment is generally willing to pay a premium to do so. Just food for thought.
eddy, i agree with you except you discount that our single family homes in the city are at all time highs as well. no reason for them to ever correct as far as i can tell too. kind of like the brownstones in manhattan. they aren’t building anymore so there’s only one way for those prices to go.
Dude,
If you are looking at a particular unit, and decide that it is cheaper to rent than to buy, that’s fine.
But if you look at “median price” and “average rent” and decide it is cheaper to rent, well, I guess some people will rent forever.
I have repeated my points several times – it makes sense to compare the price/rent for the same unit, and do a calculation of the income/price of a particular sale. It does not make sense to use average, median or whatever.
I can tell you, some of the properties on the market today are cheaper to buy than to rent for the people they are intended for (high income bracket of AMT).
Another observation – for some people, it may never make sense to buy, simply because their marginal tax bracket is low.
It is quite possible the both buyers and renters are right.
The next related question is, what happens to RE market after 2008, when we get a nice big tax hike?
I think John is on to something: real estate is probably the most tax favored investment, and that is going to continue. All taxes are only going up, and once the AMT is gone, the mortgage interest deduction in our expensive city will get even more valuable to taxpayers and may impact the p/r ratios in a way that’s difficult to account for now.
However, rent control means that it will always be cheaper for some to rent, assuming they want the same living standard. I know a lot of people in sweet Russian Hill/Pac Heights apartments who have lived there for a long time and “can’t afford” to move because buying a similar (rent controlled) apartment would cost 3-4x renting.
The mortgage deduction is capped at a million dollars. While the deduction for that million will increase if tax rates go up, there is a limit to any plausible impact of tax rates on real estate prices given SF’s already high prices — already over a million for the types of places those in the highest tax brackets are buying.
James wrote: “they aren’t building anymore [SFRs] so there’s only one way for those prices to go.” You’re only focusing on the supply side of the equation and ignoring demand. They “weren’t building any more” SFRs in 1991 either, and prices nevertheless dropped considerably for a number of years because demand dried up.
The tax savings isn’t much in the case of high end homes. The mortgage tax limits are $500,000 and $1 million (married filing jointly) for the first and second home. That is the acquired mortgage debt. Then there is another $100,000 limit on the equity line. Interest on mortgage debt over that limit is not tax deductible.
All compelling arguments…let me digest them all.
First, SF is like New York, where trees can grow to the moon because they’re not making anymore brownstones. Except said article also predicts a 34% correction in New York City. And Manhattan real estate prices have corrected before:
http://njrereport.com/80sbubble.htm
Maybe the Pacific cypress is the only tree that grows to the moon?
Next, there are apparently places in the city where buying is more economical than renting. I’m not discrediting this, I just haven’t seen any, that’s all. Could somebody post a few links to listings?
And regarding the tax advantages of buying, not arguing there. But a big chunk of this advantage gets eaten up by property taxes and HOAs. Any remaining tax advantage usually isn’t enough to move the needle in favor of owning in financial terms.
Vox, you have it all wrong, as wallstreeter pointed out. In the stock market, there will always be periods of time where the P/E of the S&P 500 is far higher than historical norms. And in the real estate market, there will always be times when the P/R is far higher than historical norms.
But just as in the stock market, as Schiller points out, these periods are also likewise followed by “reversion to the mean”.
Schiller does not say that prices can’t go up, and he’s not saying it has to correct itself tomorrow. Only that it will eventually correct itself. He only says that the difference between prices and rents will revert back to their normal relationship eventually, and as dude points out, you must either buy or rent. If it no longer makes economic sense to buy, people will not buy, they will rent. And that will cause the relationship to correct itself again.
We had a separate thread about this a week or so ago, and I posted a link to a lecture at Humboldt State from Youtube. In it, the lecturer shows how the real estate market is clearly in a bubble.
When “non-believers” hear bubble, but then see continued price increases, they say “See, there is no bubble.” But Thornberg explains that that thinking is completely wrong.
A bubble is essentially defined as a period where people are paying irrationally high prices for something. When that happens, all economic models break down because the models assume rational behavior. So economists can NEVER predict when a bubble will burst because they can’t predict irrational behavior. They can only warn that it will eventually burst because eventually people will begin behaving rationally again.
Eventually too many people will say “why pay these prices when I can rent for less?”
The reason why they haven’t said that in the past 7 years is because too many people expected the consistently high price increases to continue. One could say paying a high P/R is justified as long as the prices keep increasing by 10%. Google’s ludicrous P/E is justified if it keeps up the ludicrous growth rates.
But if those price increases halt, many buyers will say that buying a house and putting up with the additional risks, costs and burdens is not worth it when renting is so comparatively cheaper.
What will this do when it happens? It will increase demand for rentals and decrease demand for purchases, and eventually it will revert to the mean in the P/R.
It’s happened time and time again in every market in the world, whether real estate now or stocks in the late 90s or whatever. And every time, someone believes we’ve achieved some new paradigm which justifies the higher ratios. And every time in the past, those people have been proven to be wrong.
It’s perfectly fine to predict that this ratio will last for a number of years before it corrects itself. But if you’re betting we will never see the old P/R ratios again, you are making a bet against decades of economic history.
timkell,
Nicely put.
Property tax is also decuctible if you itemize on the Federal. Buying makes fiscal sense for a lot of people still, particularly for the self-employed who plan to stay in one place for a long time.
“So economists can NEVER predict when a bubble will burst because they can’t predict irrational behavior. They can only warn that it will eventually burst because eventually people will begin behaving rationally again.”
Or as The Economist magazine puts it, “economic bubbles always last longer than people expect, but they always eventually burst”.
@Dude
“vox – what are your thoughts on the Eichholtz study?”
Sure. A lot of his work says the same thing I do: for any time series one could compute a variance. That variance will underestimate volatility because it’s possible that the future will print values formerly unseen. Only if the probability distribution is normal; and, the historically observed data has comprehensively swept out the tails of that distribution, could one infer that humanity had already seen it all, and thus predict that we were looking at the ceiling or floor or somewhere in between.
@Wallstreeter
“Vox, What you are describing is a standard fluctuation which is the entire point of taking 15 years worth of data. If the study only took the last 5 years worth of data or took 5 years of data from 90 to 95 then you would have a point.”
The point would remain even if Schiller used 500 years of data. The weakness of any rolling average is that it has no predictive power because the data of any real estate price series aren’t normally distributed. Even if they were normally distributed, then contrary to Schiller’s argument for mean reversion, there is no magical numerical rubber band pulling prices back to the mean. Instead the mean is just as likely to change as the price. For further demonstration of the absurdity of mean reversion, why not create a moving average of the moving average and argue that the mean itself will then chase the mean of the historical means. Aside from misplacing the statistics of normal distributions into a domain where they don’t belong, Schiller is confusing an effect of statistical manipulation for a cause.
“If we follow your conclusions then the PE for any given stock will only go UP UP and AWAY!”
No. That’s not what’s concluded. My conclusion is that regardless of historical observation, the PE for any stock could be higher or lower than any PE formerly observed. See problem of induction.
“Shoot.. by that logic then at any PE a stock is a buy because it will keep going up. Also, any price for a home would be a bargain because it doesn’t matter how little people are paying for shelter when they don’t own a place (rent) because there will be infinite demand regardless of price for owning.”
No. That’s not logical, least of all mine.
@John
“In statistics, we would plot the data and try to figure the mean, and the standard deviation. The mean itself is meaningless. As long as the data is within one or two standard deviation, it is considered normal.”
Sure. Assuming the distribution itself is normal. How about when we observe a result that is ten or twenty standard deviations from the historical (however defined) mean? Rather than recalculate a new mean and standard deviation, perhaps it would be better to challenge the assumption that the probability distribution is normal. Man doesn’t live by one curve alone.
@timkell
No, I don’t have it all wrong. Schiller’s argument is grounded in several fallacies.
It’s a reprise of the same statistics he used to get Greenspan going on the ‘irrational exuberance’ train back in ’96. Of course, from ’96 to date, the markets far exceeded his prediction of a long term decline. He had the direction wrong.
Luckily for his speaking fees, he took the same opinion to a retail audience four years later after the markets had shredded his forecast. Even then, most indices are above their 2000 levels.
See: frequency of the accuracy of stopped clocks.
” Even if they were normally distributed, then contrary to Schiller’s argument for mean reversion, there is no magical numerical rubber band pulling prices back to the mean. Instead the mean is just as likely to change as the price. ”
Say what? I guess all the losses all the banks are booking due to losses from bad loans has no influence on price.
All you statistical mumbo jumbo is ignoring the reason for the inflated prices. The problem with just using historical data is that we haven’t had a national bubble of this magnitude caused by these conditions before. Those conditions are unwinding every day and will provide the magical rubber band that will cause the mean reversion.
@Vox,
You do have it completely wrong. Schiller predicted the stock bubble and its eventual crash correctly. Schiller did not anticipate that Greenspan’s overzealous rate reductions coupled with ridiculous lending standards would create a new bubble. The new credit and housing bubble got the markets in an irrational state again. Just look at the indices and their wild swings since august.
@Vox
You’d be correct if they were plotting real estate prices. They are not. They are plotting the P:E ratio that exists for home values. That relationship has peaks and valleys that are subject to nominal economic fluctuations, so I’d argue the distrubition in terms of how long variances last are normal distributions.
Much of this discussion exceeds my dim recollection of grad school statistics. What I do know is the following:
I rent in a small building of all rental condos. The monthly rent is about $3.68/SF. Taking into account all ownership costs (HOA fees, interior maintenance), tax deductions, equity build-up, and a normalized rate of appreciation, breakeven for me would be a price of about 200x the monthly rent ($736/SF). I would love to buy this unit…but, realistically, it would cost 275x to 300x to buy a condo similar to this one.
At current levels of rent and sales prices, renting my apartment is between 66% and 73% of the cost of owning. This gap could narrow if rents increase sharply or home prices decline by a meaningful amount.
Granted, I may have a more utilitarian view of homeownership than many. But at some point, people evaluate costs of substitutes. If one good becomes too expensive, they make choices. If you don’t like the term “reversion to the mean” then don’t call it that.
I’m not saying that prices have to fall, but they are definitely looking heavier than they did a year ago. It doesn’t take an expert in statistics to recognize that financial markets have shown patterns of excesses in both directions for as long as they have existed. It’s human nature.
@vox
I guess the P in P:E ratios in this article doesn’t stand for real estate prices then?
“That relationship has peaks and valleys that are subject to nominal economic fluctuations, so I’d argue the distrubition in terms of how long variances last are normal distributions. ”
What on earth are “nominal economic fluctuations”? Won’t nominal fluctuations affect rent and house prices the same. A unit of currency has the same purchase power whether it is used for rent or buying a house in the same time period.
Copper,
Sorry responded to the wrong person.
@akrosdabay
“Say what? I guess all the losses all the banks are booking due to losses from bad loans has no influence on price.”
That’s not inferred from anything I’ve posted. Contrarily, the issue in the CDO/SIV/XYZ markets is that prices aren’t/didn’t form within the collars predicted for these securities by any normally distributed or mean reverting models that used historical data as a precedent to predict the quality/liquidity/value of these securities. Thank you for making my point.
“All you statistical mumbo jumbo is ignoring hese secuthe reason for the inflated prices. The problem with just using historical data is that we haven’t had a national bubble of this magnitude caused by these conditions before. Those conditions are unwinding every day and will provide the magical rubber band that will cause the mean reversion.”
These “problems” are only seen as such by those who assumed that prices must revert to some mean as determined by analyzing some X years of prior history. If mean reversion actually had any power then we couldn’t have had an unprecedented bubble in the first place.
Another poster suggested that I’m arguing against decades of history. That’s not true. Decades of history tend to show that for any valuation metric, over time, there will yet be an increase in its highest and lowest values over time.
@akrosdabay
“You do have it completely wrong. Schiller predicted the stock bubble and its eventual crash correctly. Schiller did not anticipate that Greenspan’s overzealous rate reductions coupled with ridiculous lending standards would create a new bubble. The new credit and housing bubble got the markets in an irrational state again. Just look at the indices and their wild swings since august.”
No. I have it right. It cannot be disputed that Schiller mistimed the market. Had one followed the suggestion that the equity valuations of ’96 were irrationally exuberant (as suggested by the “insight” of then era mean reversion according to Schiller and his opinion of those ’96 era valuations) one would’ve faced some embarrassingly large margin calls until “proven” right, only to lose more since regardless of what US index you choose. That’s not arguable.
@Cooper
“You’d be correct if they were plotting real estate prices. They are not. They are plotting the P:E ratio that exists for home values. That relationship has peaks and valleys that are subject to nominal economic fluctuations, so I’d argue the distrubition in terms of how long variances last are normal distributions.”
You could argue that all you want. Yet, you didn’t show any data to back up your assumption of a normal distribution regarding the distribution of variance or prices.
Even assuming your conclusion, then valuations would live within set collars as already laid down by history (perhaps of the last 15 years). They haven’t, therefore any frequentist school of statisticians has little power to explain, let alone predict price action. It would be more humble if they said it in Bayesian terms: given X volatility/elasticity/etc. the probability that prices are high is X-X’%.
@vox
Schiller never claimed to or claims to time the market. You can put words in his mouth but that doesn’t make you right.
You might have a bone to pick with Schiller but you are wrong. That’s not arguable.
@vox
“These “problems” are only seen as such by those who assumed that prices must revert to some mean as determined by analyzing some X years of prior history. If mean reversion actually had any power then we couldn’t have had an unprecedented bubble in the first place.”
Rubbish. Why do we also have record foreclosures then? Shouldn’t happen if house prices only go up and never revert to the mean.
@akrosdabay
“I guess the P in P:E ratios in this article doesn’t stand for real estate prices then?
‘That relationship has peaks and valleys that are subject to nominal economic fluctuations, so I’d argue the distrubition in terms of how long variances last are normal distributions. ‘
What on earth are “nominal economic fluctuations”? Won’t nominal fluctuations affect rent and house prices the same. A unit of currency has the same purchase power whether it is used for rent or buying a house in the same time period.”
You’re quoting Cooper and attributing the quote to me. Obviously I’m skeptical about the predictive value of “nominal” fluctuations.
“You’re quoting Cooper and attributing the quote to me. Obviously I’m skeptical about the predictive value of “nominal” fluctuations.”
Already apologized for it.
“That’s not inferred from anything I’ve posted. Contrarily, the issue in the CDO/SIV/XYZ markets is that prices aren’t/didn’t form within the collars predicted for these securities by any normally distributed or mean reverting models that used historical data as a precedent to predict the quality/liquidity/value of these securities. Thank you for making my point.”
No you made my point. The reasons the models didn’t work is because the reasons for the high valuations are unprecedented and based on down right fraud. Take away those reasons and we return to the mean.
@akrosdabay
“Schiller never claimed to or claims to time the market. You can put words in his mouth but that doesn’t make you right.”
No. I don’t need to put words in his mouth ex post. Ex ante, Schiller claimed that according to the mean reverting properties of diverse data samples that prices would decline for certain securities. Schiller happened to be wrong.
“You might have a bone to pick with Schiller but you are wrong. That’s not arguable”
Contrary to your suggestion, eventually, a la Schiller, were one to assume some infinitely long data set, the mean reversion doctrine just might be recovered to some semblance of potency. This would be about as useful as saying that the US government has an annoying habit of redeeming its bonds at par, thus presenting the illusion of risklessness in owning its securities.
“No. I don’t need to put words in his mouth ex post. Ex ante, Schiller claimed that according to the mean reverting properties of diverse data samples that prices would decline for certain securities. Schiller happened to be wrong. ”
Elaborate on those securities. Unless we are residing in different worlds the stock market did crash horribly in 2001, no?
Let’s look at some facts shall we.
NASDAQ has still to return to the peak set in 2000.
The DOW has a peak of 11300 in may 2001 after which it dropped and didn’t attain 11300+ till Apr. 2006. Looks like schiller was right. 5 Years of a lull for the DOW.
The only person that is wrong here is you.
@akrosdabay
“Rubbish. Why do we also have record foreclosures then? Shouldn’t happen if house prices only go up and never revert to the mean.”
Nah. If owners of mortgage backed securities also assumed too much normalcy in debt service/volatility/etc. then they could make errors assessing their credit risk.
“The reasons the models didn’t work is because the reasons for the high valuations are unprecedented and based on down right fraud. Take away those reasons and we return to the mean.”
The point is that any analysis a la Schiller assumes that “unprecedented” valuations can’t happen in the first place because the historical sample has already shown what’s possible. Asserting that the historical sample would’ve frustrated the presently outlying modern experience were it not for fraud is two different kinds of naïve.
“The point is that any analysis a la Schiller assumes that “unprecedented” valuations can’t happen in the first place because the historical sample has already shown what’s possible.”
Not they don’t.
” Asserting that the historical sample would’ve frustrated the presently outlying modern experience were it not for fraud is two different kinds of naïve. ”
That’s the most naive statement of them all.
@akrosdabay
“Elaborate on those securities. Unless we are residing in different worlds the stock market did crash horribly in 2001, no?
Let’s look at some facts shall we.
NASDAQ has still to return to the peak set in 2000.
The DOW has a peak of 11300 in may 2001 after which it dropped and didn’t attain 11300+ till Apr. 2006. Looks like schiller was right. 5 Years of a lull for the DOW.
The only person that is wrong here is you.”
Except for the part where any person arguing for mean reversion and lower valuations (as Schiller did, as Greenspan did) in the mid ’90s on the basis of ‘irrational exuberance’ wound up getting horribly spanked by the price action ever since on any index you care to name.
The “horrible crash” of 2001 never returned any US equity index to the Schiller/Greenspan Sell Wire levels of 1996 despite the jawboning power of the world’s biggest central banker.
Sciller’s “five years of lull” call would’ve been more persuasive if they hadn’t occurred after markets already rose sharply when he said they’d fall.
all of these mathematical attempts to understand or predict the market may be appropriate for stocks and bonds, but we are talking about home ownership. For people that are willing to take a risk on RE, there is the emotional aspect to consider as well. How does a P/E ratio quantify this? All of these psuedo-economic and half-backed actuarial attempts covered in vague and complex language don’t mean anything. One can get carried away with statistical analysis, but that has limited predictability, regardless of how sophisticated. No formula can predict behavior, too many variables to consider in the real world, mostly serves to observe in hindsite. Great attempts no doubt, but purely academic, elitist, and useless. Any of you predicted that prices would double in the next 5 to 6 years after April 2000 when so much capital was lost and people were losing thier jobs in the bay area? Doubt it or you would all be home owners right now. Give these meaningless half-witted arguments a rest.
“Sciller’s “five years of lull” call would’ve been more persuasive if they hadn’t occurred after markets already rose sharply when he said they’d fall.”
Now you are grasping at straws. Please provide some evidence that Schiller said they would fall precisely at a particular date/month.
all of these mathematical attempts to understand or predict the market may be appropriate for stocks and bonds, but we are talking about home ownership. For people that are willing to take a risk on RE, there is the emotional aspect to consider as well. How does a P/E ratio quantify this? All of these psuedo-economic and half-backed actuarial attempts covered in vague and complex language don’t mean anything. One can get carried away with statistical analysis, but that has limited predictability, regardless of how sophisticated. No formula can predict behavior, too many variables to consider in the real world, mostly serves to observe in hindsite. Great attempts no doubt, but purely academic, elitist, and useless. Any of you predicted that prices would double in the next 5 to 6 years after April 2000 when so much capital was lost and people were losing thier jobs in the bay area? Doubt it or you would all be home owners right now. Give these meaningless half-witted arguments a rest.
“The “horrible crash” of 2001 never returned any US equity index to the Schiller/Greenspan Sell Wire levels of 1996 despite the jawboning power of the world’s biggest central banker.”
Really? Look at the NASDAQ.
“Except for the part where any person arguing for mean reversion and lower valuations (as Schiller did, as Greenspan did) in the mid ’90s on the basis of ‘irrational exuberance’ wound up getting horribly spanked by the price action ever since on any index you care to name.”
Except in 2001-2005 when people that listened to you probably wanted to run after you with a meat cleaver.
@akrosdabay
‘”The point is that any analysis a la Schiller assumes that “unprecedented” valuations can’t happen in the first place because the historical sample has already shown what’s possible.”
Not they don’t.’
If they don’t then there is no such phenomenon as mean reversion. Thanks for reinforcing my point.
‘” Asserting that the historical sample would’ve frustrated the presently outlying modern experience were it not for fraud is two different kinds of naïve. ”
That’s the most naive statement of them all.’
Hardly, either Schiller et. al. per mean reversion can tell us just where is the max and min valuations for Price/Rent or he is wanking himself. If the system permits events of any kind that violate the bounds of some historically normal range of values, then the range isn’t statistically normal by definition.
Similar wanking is reserved for some argument that Schiller couldn’t know what the central bank would do post 2000 (because that too was outside the bounds of prior experience?).
“If they don’t then there is no such phenomenon as mean reversion. Thanks for reinforcing my point.”
Twisting words again. I meant they don’t assume “unprecedented” valuations can’t happen.
“Similar wanking is reserved for some argument that Schiller couldn’t know what the central bank would do post 2000 (because that too was outside the bounds of prior experience?).”
I would say you have already wanked your self silly. What the central bank would do is one thing? Knowing how much of it they would do is another thing.
This discussion has reached diminishing returns.
@viewlover
Couldn’t agree with you more? No amount of statistical wankery is going to predict human emotions and behavior.
I guess that is why it’s called “irrational exuberance” and not “outside the bounds normal distribution”.
vox fails to understand this point entirely.
I am no economist, but my back-of-envelope math assures me that, with tax deductions and the like, a five-year term of living in a non-rent-controlled one bedroom high-end* apartment will be outperformed by the purchase of the same apartment as condo, if rents keep going up like they have been (or get worse, as many are predicting and I’ve already seen this fall).
(* I have no choice but to live in fairly luxurious housing, as the cheaper and older stuff with rent control is rarely wheelchair-accessible. Ask me how well Social Security pays the rent, I dare you.)
My parents decided that investing in SF property was probably about as safe and useful as the stock market, and has the added bonus of making sure their only kid doesn’t end up homeless. They’re reasonably sound investors, and their accountant didn’t make That Face, so I assume this plan isn’t throwing money away.
I’ve also seen a lot of units being taken off the sales market and rented instead “until the market comes back up”. Those rents are high. Especially the “built as condos – we’re only renting until we can sell them for more” places.
I see a lot of economic theory here about how people will just move away if housing prices get too high, but for the average low-income renter, that’s not much of an option. They may be dependent on the SF pubtrans system to get to work, or may not be able to drive at all. They may not have the spare cash on hand to pay for a move (whether that’s several thou for a moving truck, or fifty bucks of beer and pizza for the friend with a van). They may be beneficiaries of some “if you live or work in the city” health plan or other bonus which they can’t afford to give up. They may pay more in gas money or bus fare to commute from the burbs than they’d save on rent. And really, you don’t save that much in rent by leaving the city proper around here.
So people stay. Admittedly most of the poorest renters are in rent-controlled properties (and often get trapped there because they can’t afford to move out, no matter how bad conditions get.) But there’s a reason that most cities have a large contingent of working poor; somebody’s gotta make the coffee and mop the floor, and at nine bucks an hour, a long commute isn’t a money-saver. Not to mention the waiter-students, barista-opera singers and other working poor who are choosing to be city dwellers in order to work towards career goals. Most of those people pay a third to half of their income toward rent already, and will pay more if they have to in order to stay where the jobs are. Paying half your Starbucks salary for rent in the city isn’t nearly as bad as having no salary to pay rent with in Salinas, and the jobs market is still pretty painful in those brackets – lots of “underemployed” folks competing for the same jobs.
So I don’t expect prices will landslide downwards. The properties which aren’t selling will just be rented instead, at extortionate rates which will easily cover the mortgage and so on, and the property will be banked to sell again “when the market recovers”. Inventory will stay low because somebody will always rent a house in SF; this isn’t one of those places that will turn into a foreclosure ghost town. You may no longer have a million-dollar sale, but you still have a very valuable rental investment property.
Nice post wheelchairgirl. It is nice and refreshing to see something that makes sense. Vox seems to know just enough about statistics to throw together some terms that sound impressive but are completely off base.
The original post theorizes that San Francisco real estate will only correct by 10% over a 5 year timeframe. This is really not that much and over such a long time frame would even ‘appear’ to be flat. The difference will only be that someone who owns will not see crazy double digit price increases every year.
I agree with wheelchairgirl though that owning a place and being secure and accessible (or whatever you needs are) more than makes up for the 10% correction.
viewlover, you are hilarious! what about the women? the real challenge is to buy a place that acts like a fast car when they see it.
😉
does anyone stop deducting interest on a mortgage that is over 1mm? i’m just curious. i keep seeing that referenced here but i don’t see how anyone would care or know or enforce it.
Maybe we should discuss something that everybody can agree on…like One Rincon vs. Infinity.
“So I don’t expect prices will landslide downwards. The properties which aren’t selling will just be rented instead, at extortionate rates which will easily cover the mortgage and so on, and the property will be banked to sell again “when the market recovers”. ”
So this new inventory in the rental market will have no affect on rents. So rents in SF have no limits?
I guess the rents never went down after the dot-com bust in SF. People never really moved out of the city?
Are Sf residents living in a bubble of their own imagination?
“I agree with wheelchairgirl though that owning a place and being secure and accessible (or whatever you needs are) more than makes up for the 10% correction.”
A 10% correction over 5 years in nominal term destroys 50% of the initial investment ( assuming 20% down). Not counting inflation and the expense of acquiring said property which will make this only worse.
” Paying half your Starbucks salary for rent in the city isn’t nearly as bad as having no salary to pay rent with in Salinas, and the jobs market is still pretty painful in those brackets – lots of “underemployed” folks competing for the same jobs.”
I guess there is no other place to live other than “The City” in the bay area. Either live in the poorest part or put half the pay in rent.
There are these things called BART and Caltrain. Which cost about $200/ month if you commute daily. The areas outside SF which these service have rents 1/3 the rents in the city.
I don’t see very many low income families choosing to not rent a 2 bedroom apartment for half of what a studio costs in SF, that too with free parking. Not to mention very good public schools. That extra they save by not renting in SF can easily pay for the family car or be invested.
If the rents are increased to cover current mortgages expect people to leave in droves especially if incomes don’t keep up.
@wheelchairgirl
You have a unique set of circumstances and I have no doubt for you it makes more sense to buy in the city. I just don’t see others making the same set of calculations based on their circumstances.
Wheelchair girl makes a very good point that there are many reasons to buy rather than rent that are outside of financial considerations — like her parents making sure their only kid doesn’t end up homeless. But she needs to check her back of the envelope math — unless she is assuming extremely high rent escalation over the next five years (or extremely low market returns), from a purely financial standpoint there is no way that buying a condo at today’s prices is less expensive than renting the same condo. That doesn’t mean buying never makes sense, but given the near universal consensus that prices aren’t going up and are likely to come down over the next few years at least, buying is not a smart financial move right now (unless you are looking at a very long period of ownership, which is rare these days) — and buying a place to rent it out is even worse.
Here is the NY Times rent vs. own calculator — previously posted here and very good:
http://www.nytimes.com/2007/04/10/business/2007_BUYRENT_GRAPHIC.html?_r=1&oref=slogin
“Except for the part where any person arguing for mean reversion and lower valuations (as Schiller did, as Greenspan did) in the mid ’90s on the basis of ‘irrational exuberance’ wound up getting horribly spanked by the price action ever since on any index you care to name.”
Schiller is always very careful not to argue when valuations will go down, because part of his thesis is that bubbles are a mania and thus take a surprisingly long time to burst. His overriding point is that bubbles do eventually do all burst.
The argument seems to be that, since the smartest economists can’t predict the timing of a bubble-burst then bubbles can’t exist. Only hindsight will tell us if we experienced an unsustainable price increase over the past several years, or if this is a “permanently high plateau” in real estate prices. Neither the housing bulls or bears can make a completely persuasive case.
A deep and prolonged downturn may or may not happen. If it does, many homeowners won’t care because they don’t look at their home as a financial asset and they don’t plan to move anytime soon. However, it’s worthwhile for everyone to consider how they would be affected by various “black swan events” and, if necessary, take precautions. A little insurance isn’t a bad thing.
A few observations and gloom and doom predictions about the housing market:
1.) Investors who were huge buyers of CDOs are getting killed and consequently are no longer buying them. Without buyers of CDOs, Wall Street can no longer buy mortgages and repackage them into CDOs to sell to investors. Because Wall Street is no longer buying mortgages, mortgage lenders cannot originate mortgages unless they can sell them to Fannie Mae or Freddie Mac (which means the loans must conform to less than $417,000 in principal, no subprime, and more important to bay area buyers, NO JUMBO). Anyone who thinks this is just a subprime problem does not understand the mortgage market. Until Wall Street has cleaned up its balance sheet and taken all the losses it needs to take, the situation will not get any better.
2.) The run up in housing prices all over the country from 2003-2005/6 was a financing phenomenon. The financing spigot has now been closed. Unless you have extremely good credit, have all cash, or are purchasing a house for under $417,000, your financing options have greatly decreased.
3.) To those on this website who wonder why Manhattan housing prices have continued to explode despite the rest of the country’s real estate price deflation, the explanation is this: $36 billion in bonuses paid out last year to Wall Streeters. Given the huge losses all the banks are taking now, many who received bonuses last year will be getting pink slips this year and will be looking to puke their $2 million condos. Manhattan real estate prices are headed lower in the New Year.
4.) As for SF, while it is true that the job market here remains strong because of tech, and a lack of housing supply, the financing problems mentioned above will affect prices here as well. It is already apparent in the increased housing inventory numbers and the stagnant prices of the past few months. Also, a casual flip through the real estate section of the Chronicle shows new home builders holding auctions of tons of properties they are looking to unload in the Bay Area as well as ads for upcoming foreclosure auctions. San Francisco does not live in a bubble, unless of course, you mean financial bubble…
Feel free to heckle me in the coming months if I am wrong.
@ predictiongirl
Your arguments would hold a lot more water if you bolstered them with some arcane statistical theory 🙂
“The argument seems to be that, since the smartest economists can’t predict the timing of a bubble-burst then bubbles can’t exist. Only hindsight will tell us if we experienced an unsustainable price increase over the past several years, or if this is a “permanently high plateau” in real estate prices. Neither the housing bulls or bears can make a completely persuasive case.”
Indeed, because the world is rarely such a black and white place that the future can be predicted with certainty. But the timing argument seems akin to saying that because the best scientists in the world can’t predict when the next big quake on the Hayward Fault will be, we can’t be sure that stress is building on the fault or whether a new “permanently high plateau” of stress exists that won’t cause a quake. Good luck if you think the second is true.
We’re now repeating and missing earlier parts of the conversation.
Vox, either you are still not grasping the difference between a data point, housing prices, and a ratio, P/R, or P/E, or you are just refusing to do so in order to continue arguing.
In the stock market collapse, stock prices did not need to drop down to 1996 levels to make Schiller right, as you argued above. P/Es needed to come back inline with historic norms. From 1996 to 2001, the economy experienced amazing growth. That means the “E” in P/E is higher, which means the “P” must be higher. Schiller predicted it would happen, and he was right.
Same goes for housing. Schiller is not saying the “P” must drop. Only that the P/R must drop. And it must drop eventually, either through increased rents or decreased prices or a combination.
Also just because there’s a chance that at some point in the future P/Rs might be higher than they are today, or lower than they’ve ever been before, that has nothing to do with Schiller’s point. Even when those times occur, one can still safely bet that eventually, the ratio will return to its mean.
Further, as I said earlier, economists cannot and will not ever be able to predict the end of a bubble. The very definition of a bubble is that people are paying irrationally high prices for something. Economics is based on an assumption of rational behavior. You can’t do modeling and predicting when people are behaving irrationally.
So Schiller can hazard a guess as to the timing of a correction, but it’s only a guess. He is right to say that eventually the P/R will return to a more normal ratio, but it would be difficult to predict exactly when that will happen.
I could be a bit wrong in the case of real estate because real estate bubbles take much longer to unwind than stock market bubbles. Schiller could believe he’s facing clear evidence that the bubble has indeed burst, and he’s using modeling to predict the length of time it will take to unwind.
Due to the vastly greater liquidity of the stock market, a bubble can burst and be overdone in one day sometimes. Not the case for real estate.
Link to the humboldt lecture again:
http://www.youtube.com/watch?v=uyOWuczlJCA
Amen Corner
“The argument seems to be that, since the smartest economists can’t predict the timing of a bubble-burst then bubbles can’t exist.”
This was actually intended as a snark against the “anti-bubbleonians”. As I suggested in the final sentence of that comment, it might be best not to build on your home on a fault-line or your entire financial future on a risky bet if you couldn’t live with the wrong outcome.
End of part 1 of that lecture, Thornberg shows how at the end of 2003, all the fundamentals reversed. Mortgage rates had nowhere to go but up. Builders were building 1.8 units per new household, increasing available inventory. Rents stagnated, incomes stagnated.
Fundamentals all stagnated or reversed, which should have triggered housing to stagnate. Instead, prices still continued to explode. When prices substantially deviate from their underlying fundamentals, that’s when you have a bubble.
Oh yeah, he also mentions last 500 years real (meaning adjusted for infaltion) price increases in housing in Denmark. Guess what is? 0.1%. How about US history? 0.2%. Last seven years before ’07? ~10%? Interesting.
From part 2:
“It HAD to end. Figuring it WHEN it will end is difficult… Asking a forecaster when a bubble is going to end is like asking a psychologist what a crazy person is going to say next.:
LOL.
“But at some point in time, there was going to come a time when someone was going to wake up and say ‘My God, you’re asking $900,000 for a 2BR 1BA apt conversion in 17th St. on Santa Monica in the basement, with a view of somebody’s tailpipe. You must be insane.’ As soon as that happens, it’s over.”
Haha.
i only listen to audio lectures form harvard or yale.
humboldt is known for weed, not academia
Here’s Schiller in July of 1996 (see: http://www.econ.yale.edu/~shiller/data/peratio.html)
“Looking at the diagram, it is hard to come away without a feeling that the market is quite likely to decline substantially in value over the succeeding ten years; it appears that long run investors should stay out of the market for the next decade…The fitted value for today of the regression is –.479, implying an expected decline in the real Standard and Poor Index over the next 10 years of 38.07%.”
See those nice graphs of P/E ratios labeled “P/E ratio predicts later real P change”? Ha!
He then hedges a bit to show that his most optimistic expectation is no return at all.
Finally he admits that he could simply be waving arms, in which case all bets are off:
“The conclusion of this paper that the stock market is expected to decline over the next ten ears and to earn a total return of just about nothing has to be interpreted with great caution.
“Our search over economic relations that are used to study the price divided by 30-year moving average of earnings may have stumbled upon a chance relation with no significance. In other words, the relation studied here might be a spurious relation, the result of data mining. Neither the statistical tests nor the monte carlo experiments take account of the search over other possible relations.
It is also dangerous to assume that historical relations are necessarily applicable to the future. There could be fundamental structural changes occurring now that mean that the past of the stock market is no longer a guide to the future.”
In layman’s terms, Schiller says, “According to my analysis the stock market is about to go into the toilet for a very long time unless my analysis is just a lot of meaningless wankery in front of a rear view mirror”.
So how did this prediction work out for the good Professor? Looking at HIS chosen index (the S&P) in NO month for the next ten years thereafter was the market ever lower than its 639.95 value at the end of July 2006. He was NEVER right over any time scale! One year later the S&P was at 954.31. Over ten years, without even accounting for the dividends and dividend reinvestment that would push up total returns, the index increased 99.49% for a compound annual growth rate of 7.15%.
“Chance relation”,”no significance”,”spurious relation”,”could be fundamental structural changes”… Schiller’s own out clause makes all my points as to why the same type of analysis of Price/Rent in 2007 is as full of FAIL as his analysis of the S&P in 1996 proved to be.
Bad economist! No regression curve for you.
I was tempted not to respond to your nonsense.
“Here’s Schiller in July of 1996 (see: http://www.econ.yale.edu/~shiller/data/peratio.html)”
Object not found! Bad armchair economist no graph for you.
”
So how did this prediction work out for the good Professor? Looking at HIS chosen index (the S&P) in NO month for the next ten years thereafter was the market ever lower than its 639.95 value at the end of July 2006. He was NEVER right over any time scale! One year later the S&P was at 954.31. Over ten years, without even accounting for the dividends and dividend reinvestment that would push up total returns, the index increased 99.49% for a compound annual growth rate of 7.15%.”
More unsubstantiated rubbish. The S&P 500 had a peak of 1552 in 2000. With the low close of 815 in 2002.
It is at 1494 today. Ding! Ding! Ding! vox is speaking from his nether regions again.
It is getting very hard to take you seriously.
http://finance.yahoo.com/charts#chart1:symbol=^oex;range=my;charttype=line;crosshair=on;logscale=on;source=undefined
Look ma! S&P 100
Peak in 2000 -> 834
Nov 2007 high -> 722
Never reached the bubble levels.
Vox sorry you are wrong again… Schiller was right.
I would have loved to have seen vox telling people in 2002 that nothing bad happened with the stock market. LOL. You’re priceless.
You are using information showing that he was off perhaps in timing and perhaps degree to show that his predictions should be completely ignored. Do so at your own peril.
@vox
It is very evident you don’t understand complex economic systems. You want to pretend two back to back bubbles haven’t existed over the last decade. That’s your prerogative.
But all your talk suggest you have blinders on. It may be in your best interests to take them off.
@vox
One more thing. Schiller does math to show real values not nominal values as you seem to have a hard on for.
Factor in the dollars depreciation and those securities might have already reverted to the mean in real terms.
If you want to go right to the horse’s mouth, Schiller also likes to massage his numbers by adjusting valuations for inflation.
See: http://www.irrationalexuberance.com/shiller_downloads/ie_data.xls
Over Schiller’s 1500+ data points, the mean PE was 16.2. In July of ’96 it was 24.86. Over the decade thereafter it rose to 44.19 and its low was 21.21. In all cases it stayed at least 31% above the mean.
Don’t want to take Schiller’s own word for it? How about with extra inflation added goodness? Same result. No mean reversion.
See: http://www.investopedia.com/articles/technical/04/020404.asp
Or: http://www.sharelynx.com/chartsfixed/Barrons/ABN.gif
And: http://tal.marketgauge.com/dvMGPro/charts/Charts.asp?chart=PERATI
Notice how in the third chart the PEs only lately began to skim the upper band of what was formerly considered the “ahem” normal range? Still a far cry from any historical mean.
Fine let’s use Schiller’s real values 1996-date.
PE at the time Schiller wrote the piece in ’96: 24.86
Peak: 44.19
Low: 21.21
Lately: 24.69
Mean on his own time series over 1,500 data points: 16.2
No mean reversion. Schiller’s real PE was never less than 30.93% above his own real mean PE.
the q1 2007 p/e for the s&p 500 was 17.09. your numbers are wrong.
@vox
Let’s see you either post broken links or articles that contradict your own statements.
“Notice how in the third chart the PEs only lately began to skim the upper band of what was formerly considered the “ahem” normal range? Still a far cry from any historical mean.”
Once again blinders on. Looks like they are reverting to the mean.
Schiller claimed stock would crash they did. Schiller claimed housing would eventually crash. Oh Look! It has been for all of the this year.
Too bad the facts always seem to disagree with you.
Wow. I can’t believe you guys argued about this til 1am last night…
@timkell
“the q1 2007 p/e for the S&P 500 was 17.09 your numbers are wrong.”
If only. Akrosdabay want’s to play with inflation adjusted numbers. These are Schiller’s own numbers.
I will confess to a certain my bad by calculation the mean based upon post ’96 numbers. When Schiller wrote his paper, the mean was then 15.48 not 16.2. That means subsequent real PEs were always at least 36.9% above the mean, according the Schiller’s own data. Thanks to subsequent and sustained increases in the PE, the real mean PE as Schiller calculates it then rose to 16.2.
And your 17.09 is still higher than any historical mean. Once upon a time any nominal PE in the 17 range was a great place to short.
“looks like they are reverting to the mean”
’cause they’re now only 61% above Schiller’s ’96 mean and once they were 185% above the mean (which happened years after Schiller said they’d revert back to mean)? Obviously this mean reversion mechanic isn’t doing it’s job. If the stock market can sneer at it’s mean for over a decade, one wonders how a Price/Rent index can fight off its measly 27.7% overvaluation.
“Get in there and sell! Where’s Beakes?”
Vox, as I said, you’re using a minor error of timing/degree to claim the entire premise is false. Never upon a time in the last 50 years was a 17 P/E a time to short unless you’re a day trader. If S&P 500 P/R is in the range of 16-18 it’s within historic standards. When it’s at 45, you have a problem. When it’s at 17.08, it’s fine. Sure, maybe the fed lowering discount rate causes a minor bump to bring it a bit, but it’s a far cry from “good time to short”. There are plenty of valid reasons for changes in the ratio at a give point in time. It’s when you’re at the extremes that you have to worry, as the stock market crash of ’00-’02 showed.
Where are you getting your numbers? They’re wrong, wherever they’re from, so stop using them:
http://www.ny.frb.org/research/directors_charts/ipage20.pdf
P/E in 1997 was in the range of 22 to 25. A bit hot for historic standards and right to be a bit concerned that it wouldn’t hold up. As of 10/31/07, it was up to 18.25 with the Fed cuts, but now it’s dropped a bit since then.
You see that 37 in ’02? It’s been a nice ride down ever since, despite the fact that we’ve had nice gains in the past year. Why? Because the E in P/E went up too.
Sorry, 47, not 37.
@timkell
“Vox, as I said, you’re using a minor error of timing/degree to claim the entire premise is false.”
Backtest that. Short an “overvalued” S&P any time from ’96 to date. It’s a wipeout, not a minor error. And no, the strategy never gave a buy/cover signal since 1996, making it consistently a wipeout for over a decade except for some recent months when it would have paid small gains.
“Never upon a time in the last 50 years was a 17 P/E a time to short unless you’re a day trader.”
Read up on what happened from the middle 60s to ’74 or ’82. Those are quite long term holding periods, not day trades. In constant ’66 dollars, the loss on the S&P index would be -54% as of ’74 and -61.3% as of ’82.
If you want to go back more than 50 years, then there are more examples of why it was once thought wise to sell at a 17 PE.
This is the essence of Schiller’s problem: from the ’90s to date, the stock market has successfully defied the nonexistent gravity of its mean valuation for a very long time.
Some historical correlation isn’t a source of future causation.
RE shorting, I meant you never short simply because it’s at 17. You may short because you expect an economic implosion, etc. of course.
“from the ’90s to date, the stock market has successfully defied the nonexistent gravity of its mean valuation for a very long time.”
Explain how going from a P/E of 45 to a P/E of 17 is defying gravity please.
The Fortune article says that the P/R ratio in SF in 6/2007 is 38. I’ve been looking at what’s currently on the MLS and estimating what the properties would rent for, and also vice versa with rental properties, and I’m consistently coming up with a P/R of 20 – 22. Does anyone else think that Fortune’s P/R is ludicrously high? Perhaps they factor in rent control to get a small denominator?
This is not directly related, but here are the latest results from my script that trolls the sfgate.com/housing data and calculates volume and averages. Looks like we’re finally starting to see the effects of the August credit crunch show up there, although it’s not nearly as pronounced as I would expect:
San Francisco:
Alameda:
Contra Costa
Again, I have no idea how valid the data is due to issues like completeness, latency/timeliness, etc. But those are the numbers.
“Backtest that. Short an “overvalued” S&P any time from ’96 to date. It’s a wipeout, not a minor error. And no, the strategy never gave a buy/cover signal since 1996, making it consistently a wipeout for over a decade except for some recent months when it would have paid small gains.”
How about from 2000 till date? Exactly!
You have this hard on for schiller and 1996. Get over it! no amount of you harping is going to convince anyone that there wasn’t a dot-com bubble that burst. Or that schiller didn’t predict it before anyone ever noticed.
@Vox
No because a P/E raio is indicates a psycological relationship. i.e. borders of behavior that people are willing to pay. The level of one doesn’t matter –it’s the ratio of the two that counts. If rents get way out of wack of income then people will naturally revert to the mean of the series. That constraint doesn’t exist in a stock price.
I think rent control does distort this number. My condo has a p/r of about 22, but someone paying $2000/month in a rent controlled Pac Heights apartment that would sell for $900K has a p/r of 45. This latter number could be applied to literally hundreds, if not thousands, of units but comparable for sale properties are pretty rare. In other words, how often do 2BR condos in Russian Hill come on the market? If the answer is not very often, then how useful is it to say that there are 1000 rent controlled apartments in the neighborhood with p/r ratios of 40-50?
It seems like this just skews the p/r ratio higher. If Gdog is correct and the typical housing unit for sale in S.F. right now has a p/r of 20-22, isn’t that far more significant that the p/r ratios for housing units that will never hit the market. How many places sold during the boom with p/r ratios of 35+? I suspect not many. Even Palms condos that recently sold for 900K are renting in the $3500 range- that’s a p/r of about 26.
I really don’t know if such suppositions are correct, but to suggest that the current p/r ratio is 38 just seems off.
gdog, are you including property taxes and estimated maintenance costs? I believe they are using TCO, not just price.
Not sure where rent comes from, but the property tax and insurance costs alone would change your ratio a decent amount.
timkell,
Price/rent definition is simple – it is the price vs rent. The factors you quoted will influence what P/R number is rational for buying, but is not factored in the calculation, no matter who does the calculation.
@Cooper
“No because a P/E raio is indicates a psycological relationship. i.e. borders of behavior that people are willing to pay. The level of one doesn’t matter –it’s the ratio of the two that counts. If rents get way out of wack of income then people will naturally revert to the mean of the series. That constraint doesn’t exist in a stock price”
I think you’re reading this upside down. If rent were high as a percentage of sale value, esp. at a level that deterred renters (or encouraged them to become owners), then the building would have a low Price/Rent ratio.
Whereas if you’d like to smoke statistical crack with Robert Schiller, he’s busy scratching his head about why landlords are now willing to rent their places for such a small percentage of their sale value: a high P/R and higher than mean P/R. According to Schiller’s thinking, rents in SF are now too low; and, why would anyone buy when they could rent? Agree with him?
Schiller’s thinking at 1/38.2 (the survey P/R) the earnings yield of those SF buildings is a measly 2.6%. But according to his trusty mean (that makes things revert to it [NOT]) landlords are used to demanding an average rent of 1/27.4 of the sale value= 3.6% earnings yield over the last 15 years.
Why are these landlords being so irrationally generous when surely they must have those mean ol’ means yanking their lease rates way higher? Or, anticipating those lusciously sweeter rents to be had, they’ll naturally rush to the exits to dump their overvalued properties and buy Beanie Babies, or somethin’. Right.
Just as in the case of the stock market, the RE market is capable of ignoring this mean and increasing the P/R: bitter renters go back home to the affordable Midwestern homes of their childhood and/or people who stay pay even higher prices to buy.
“Whereas if you’d like to smoke statistical crack with Robert Schiller, he’s busy scratching his head about why landlords are now willing to rent their places for such a small percentage of their sale value: a high P/R and higher than mean P/R. According to Schiller’s thinking, rents in SF are now too low; and, why would anyone buy when they could rent? Agree with him?”
Landords aren’t willing to rent out at low prices the markets demand it.
I agree with him.
So far you haven’t been right about much.
It’s difficult to find comparable single family homes that are for rent and for sale. That house on Holladay that was posted on Socketsite is selling for $849,000 and renting for $4000. That is a p/r ratio of only 17.7.
For a p/r of 38.2 (the number claimed by Fortune), the monthly rent on the $849,000 house would have to $1852. There is no way you could rent a new, solar powered house in Bernal Heights for $1852/month, even one overlooking 101.
@akrosdabay
“How about from 2000 till date? Exactly!”
Even if you’d like to make him so lucky as to publish and short at high tick then up until today you’re still down because the short seller has to cough up the dividends.
“You have this hard on for schiller and 1996. Get over it! no amount of you harping is going to convince anyone that there wasn’t a dot-com bubble that burst. Or that schiller didn’t predict it before anyone ever noticed.”
I don’t have a hard on for him, obviously. I do agree with his own comments about how his analysis in ’96 (and ever after) was possibly a useless bit of arm waving (it was). Nobody said there wan’t a dot-com bubble that burst– it’s merely irrelevant to the understnading of Schiller’s regression curve. But thanks for repeating the strawman.
“Just as in the case of the stock market, the RE market is capable of ignoring this mean and increasing the P/R: bitter renters go back home to the affordable Midwestern homes of their childhood and/or people who stay pay even higher prices to buy.”
Stock markets aren’t an entity. Investors are irrational.
If the stock markets are ignoring means and housing is too. Why is the DJIA down 980 pts from its 52 week high…? And the S&P 500 100pts. They should be up and up and up according to your wanking.
Watch closely in the next few year how the Stock Market and housing ignore your ramblings and you are left scratching your head.
“It’s difficult to find comparable single family homes that are for rent and for sale. That house on Holladay that was posted on Socketsite is selling for $849,000 and renting for $4000. That is a p/r ratio of only 17.7.”
Has it sold or been occupied by a renter? The rent is priced to high and so is the house. No takers means price too high.
For apts/condos:
According to
http://www.rent-sf.com/avg_rent.html
the average rent for a 1 bedroom apt in Bernal Heights is $1918. You can buy a 1 bedroom TIC in Bernal Heighs right now for under $400,000 (2 Coso Ave.) That is a p/r of 17.4.
The cheapest 2 bd condo at Mission Valencia is $599,000. The average rent for a 2 bd apt in Bernal Heights is $2517, and for the Mission is $2672. That is a p/r of 19.8 for Mission Valencia compared to apts for rent in Bernal Heights, and a p/r of 18.7 compared to apts for rent in the Mission.
“It’s difficult to find comparable single family homes that are for rent and for sale. That house on Holladay that was posted on Socketsite is selling for $849,000 and renting for $4000. That is a p/r ratio of only 17.7.”
That house will never rent for more than $2800/month
“The cheapest 2 bd condo at Mission Valencia is $599,000. The average rent for a 2 bd apt in Bernal Heights is $2517, and for the Mission is $2672. That is a p/r of 19.8 for Mission Valencia compared to apts for rent in Bernal Heights, and a p/r of 18.7 compared to apts for rent in the Mission.”
Those rental numbers are complete bollocks.
From CL, the average rent for a 2bdroom in bernal is $2075 and in Mission is $2250. Even these seem quite high to me.
+ you have to take into account property taxes + broker fees, etc.
Here’s a 1 bedroom apt right on Mission St, across the street from Mission Valencia (you can see Mission Valencia out the window in the Craiglist pic):
http://sfbay.craigslist.org/sfc/apa/471121890.html
A 1 bedroom on Mission St, and they are asking $2195/month!
Gdog, Anono
The ratios you quote seem more reasonable to me than the ones listed in the article. As I mentioned in my post 11/06 10:27 p.m., I estimate a P/R of about 23-25 for the market-rate unit that I rent. It’s still too high…I reckon that buying would be a breakeven proposition at about 16.7x annual rent (taking all financial benefits and costs of ownership into account).
Based on Dan’s numbers, I must be paying well under market rates for rent on my (not rent controlled) apartment. The thing is, when I look on craigslist, I see a lot of places that look comparable.
Take this one for example.
http://sfbay.craigslist.org/sfc/apa/467248863.html.
It’s a “Beautiful 3 Bedroom 2 Bathroom House with View” for $2875/month. It looks like a mediocre single family home on a busy street. At a p/r of 18, that house should be available for sale at $621k. You might be able to find a tear down for $621k. There is no recent sale info on this, but it does look like the owners paid $847 in property taxes last year. $847?!? Where is that prop 13 discussion?
How about this one in Nob hill?
1 Bd room and 1 care parking for $1900.
http://sfbay.craigslist.org/sfc/apa/472072368.html
There is a difference between what the asking rent is and what they will get for the place in Mission.
We can do this all day.
Heehee, to some people, both rent and price are too high.
this link is cool since it gives you a quick glance at what folks are asking in all kinds of parts of town, nice and fringe area:
http://www.sanfranciscosentinel.com/?page_id=680
If you look at the average asking rents on the website I linked to above, you can see that rents appear to be rising rapidly. For example, from the same sources, the average asking rent for a 1 bedroom in Bernal Heights was $1396 per month in 2005 and $1505/month in 2006. However, the average asking rent for the past month for a 1 bedroom in Bernal Heights is much higher: $1918/month.
@akrosdabay
“Stock markets aren’t an entity. Investors are irrational.”
Special pleading for irrational behavior doesn’t support your advocacy for mean reversion. It’s an argument for mean aversion, though.
“If the stock markets are ignoring means and housing is too. Why is the DJIA down 980 pts from its 52 week high…? And the S&P 500 100pts. They should be up and up and up according to your wanking.”
No, not at all. Once again you’re putting words in my mouth to erect yet another strawman. You’re also continuing to confuse a lack of restraint on a valuation multiple (the esence of your strawman argument that I’m arguing for prices in the stars) for a lack of restraint on the variance of data in a time series (which is why mean reversion is a waste of time because means under sufficiently fat tails aren’t meaningful).
Even if forced to live within the narrow minded collars of normal distributions, there’s a perfectly plausible answer: randomness. It’s well within the bounds of the expected short term volatility as forecast by trading in the VIX at the time of the all time highs to which you refer.
“Watch closely in the next few year how the Stock Market and housing ignore your ramblings and you are left scratching your head.”
Since your persistent erection of strawman arguments shows you didn’t understand what I was commenting about from the get go, I’ll chalk up this latest tedious and unartful ad hominem jab to your lack of familiarity with statistics.
You have to love Vox’s persistence. So far he’s not uttered one single word to convince anyone else that there’s no merits to reversion to mean in the areas of P/E and P/R, and yet he still speaks as if he’s clearly demonstrated his genius for all to see.
3 cheers for Vox.
Vox, let’s assume “mean” S&P P/E is 17 for past 20 years. Barring some fundamental change in how economies work, do you expect to see the S&P P/E ever hitting 17 again?
Or is your thesis simply, “it’s all random so why bother predicting anyway?”
@timkell
“You have to love Vox’s persistence. So far he’s not uttered one single word to convince anyone else that there’s no merits to reversion to mean in the areas of P/E and P/R, and yet he still speaks as if he’s clearly demonstrated his genius for all to see.”
Well, at least I tried. Some of you aren’t ready to kick that Gambler’s Fallacy habit.
In the stock market, the empirical reality of leptokurtotic returns (or valuation multiples) with right hand tails well beyond three standard deviations from any long term mean makes mean reversion not useful. Similarly, if you graphed the price or P/R of all housing you would find prices and rents and P/Rs for certain properties and P/Rs as a whole that were more than 3 standard deviations from a long term mean. All of the above is “unpossible” according to any application of mean reversion whose expectations of volatility and a fixed upper and lower boundary to the values of future events are anchored to an arbitrarily measured variance.
“3 cheers for Vox.
Vox, let’s assume “mean” S&P P/E is 17 for past 20 years. Barring some fundamental change in how economies work, do you expect to see the S&P P/E ever hitting 17 again?
Or is your thesis simply, ‘it’s all random so why bother predicting anyway?'”
Even the advocates of mean reversion acknowledge that over shorter time horizons predictability is frustrated by noise (randomness). Our favorite Yale economist acknowledges the same thing which is why he needs a 10 year rollforward for his predictive metric to work as he hopes.
In the middle ground are people like me who assert that fundamentals apply but they too cannot be forecasted; the randomness in the future earnings (or growth rate or dividend yield or…) and the measurement error of any mean PE makes it useless to use a fundamentalist mean reversion for predictions. Even if mean reversion did happen, mean reversionists are unable to predict the size of the residuals from their regression lines (as happened to Shiller from ’96-00 when valuations increased their distance from his regression line)so what’s the point?
Then on the other end of the spectrum are the pure efficient market theorists and their random walks.
Using existing expectations for volatility, it’s not at all hard to accept that stocks could trade at a 17 PE sometime in the future. Just today, VIX is at 26.xx so a return to a mean valuation from here is within one standard deviation (as is a move the other way).
Personally, since I think those measures of expected volatility tend to underestimate volatility, I am more relaxed about expecting moves of X% over Y years than the market is.
That’s where you lose me, though. Just because Schiller is only 90% right does not mean he should be ignored.
I agree with you it’s difficult to predict, but just because something’s difficult and you might miss the mark, that doesn’t mean you should just pretend it won’t happen. It will happen, at some point, we just don’t know when. Sure things can go up again beforehand.
It’s one data point, and history shows it will eventually be true. That doesn’t mean one should base their purchasing decisions entirely on this report or any other data point. As you rightly said, there are times when you can leave a lot of money on the table if you leave a bull or bubble market too early. But that doesn’t discount the fact it will end eventually, and if I’m considering putting my life savings into something, I think the information is useful.
It’s unfortunate, but many people have spent a lot more than they wished to buy a place. They convinced themselves it made sense because they assumed the price would continue climbing. What the report highlights is the level of risk associated with purchasing real estate. It’s not a “no brainer/can’t miss” investment opportunity, as some became convinced of in the recent past.
If you’re depending on a jump in the next two years to justify a condo purchase, you should stay out of the market. If you can handle a minor drop, hop in and enjoy your new digs.
P/R is still over 30 in msot areas of the city. period
“That’s where you lose me, though. Just because Schiller is only 90% right does not mean he should be ignored.”
Grading his irrational exuberance paper, the only grades in so far are the ten year rollforwards from ’96-’06 and ’97-’07 and those grades missed the mark. The jury’s still out on any later rollforwards.
“I agree with you it’s difficult to predict, but just because something’s difficult and you might miss the mark, that doesn’t mean you should just pretend it won’t happen. It will happen, at some point, we just don’t know when. Sure things can go up again beforehand.”
A mean reversionist believes that prices have some kind of memory of more rational days (where rational is defined as a place between high and low); some mean puts its thumb on the scales of fate and distorts the 50/50 probability of the efficient market.
To “prove” their point a post hoc analysis of any time series will show that any price will cross some mean. Which only begs the question what mean do you use? It will also show residuals that can be far away from the mean. In hindsight, the bigger residuals are called bubbles and panics.
In my camp, the ad hoc problem for the investor is that this “memory” is so amnesic and selective that is makes for an unreliable prophet. One can’t know which mean will be crossed and the short term means don’t yield a workable strategy to capture excess return. the long term means get left out in the cold for a loooonnnnngggg time. One can’t know in advance whether its the P or the E or both that’s going to effect the mean reversion. One can’t know whether the future will bear witness to an event never seen before in the past: this is the problem of induction. Cases in point: San Francisco real estate values didn’t discount a Gold Rush or an ’06 catastrophe before these events. Any land buyer of the day who had looked backwards to set their expectations of future values or volatility would’ve been very surprised.
Efficient market theorists simply reject mean reversion: subsequent prices move independently of their prior position without any memory of where they used to be. In this view, if I flipped heads 50% of the time over a 1,000 flips and then flipped 10 heads in a row, it doesn’t make the 1,011th flip more than 50% likely to be tails just so the trail of flips will revert to the 50/50 mean distribution of heads and tails.
“It’s unfortunate, but many people have spent a lot more than they wished to buy a place. They convinced themselves it made sense because they assumed the price would continue climbing. What the report highlights is the level of risk associated with purchasing real estate. It’s not a “no brainer/can’t miss” investment opportunity, as some became convinced of in the recent past.”
True. The trick is that when the P/R rises and the cap rate falls, it could also be the result of an increased risk tolerance or longer investment horizon amongst the buyers. The fun in RE is that it spends long periods in a non-trending market and shorter periods in a trending market. This cranks up the leverage because the lender is convinced by non-trending markets of the stability of prices while the borrower is convinced by trending markets that prices are unstable.
“If you’re depending on a jump in the next two years to justify a condo purchase, you should stay out of the market. If you can handle a minor drop, hop in and enjoy your new digs.”
You can never depend on a near term jump, even when prices are in a panicky toilet because they’re not usefully mean reverting, are noisy and are exposed to event risks! The converse is also true.
“You can never depend on a near term jump, even when prices are in a panicky toilet because they’re not usefully mean reverting, are noisy and are exposed to event risks! The converse is also true.”
That’s for sure. Unfortunately, so many people did just that. I’m sure we all know a couple who bought some tiny little apt, thinking “oh, we’ll just live here for two years and then we can upgrade.”
I hope you can live there for 4 or 5…
“Since your persistent erection of strawman arguments shows you didn’t understand what I was commenting about from the get go, I’ll chalk up this latest tedious and unartful ad hominem jab to your lack of familiarity with statistics.”
You should look up the meaning of “ad hominem”. Then read this thread again to see who started the ad hominem jabs.
Unfortunately for you you can’t hide your lack of understanding of the world economy today under arguments based on your knowledge of statistics. If your only tool is a hammer…..
Housing prices will revert to the mean as dictated by incomes and rents. You can wank off to your statistical inference of the data on how it won’t.
“Efficient market theorists simply reject mean reversion:”
That explains it all. You are one of those. That explains the religious fanaticism and absolute disdain for Schiller.
“Housing prices will revert to the mean as dictated by incomes and rents. You can wank off to your statistical inference of the data on how it won’t. ”
Except the mean is a moving target. If the P/R ratio stays 38 for 15 years, we will be looking at a mean of 38 in 2022. Did the ratio catch up with mean, or did the mean catch up with the ratio?
Statistics won! This 2007 predictions of housing depreciation is alarmingly accurate.