How bad is housing?

Gee, starting such a novel thread. Steve really is such a thought leader, and original thinker. And doesn't repeat the same points over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over and over.........................

Nice post, Steve. I think you bring a lot to the table here, and I think that those who disagree with you should do so on the merits, and not resort to personal attacks. Unlike many on this blog, I have owned and sold real estate for many years. In 2003 I bough into a new development downtown for around 1.5. People told me I overpaid, but I liked the location, it was 1800 feet with high ceilings, new finishes, etc. ,and the kicker was that with my 5% interest only mortgage, my after tax expenses more or less equaled what I was paying in rent for an inferior unit. For the last several years I have been bullish on Manhattan real estate, and I ignored those who,in 05, 06 and 07 claimed that the market was crashing because, in my assessment, the fundamentals were strong.
In late 07 I began debating Steve on where real estate was going, and I was on the opposite side of the argument from him. However, after analyzing his arguments, as well as my own thoughts on the rapidly deteriorating financial markets, I decided that it was time to get out. While it was true that I was moving, I did not need the proceeds of the sale for my new house I could have rented or simply held onto the place and use it a place to stay in Manhattan on late nights in the city. However,I became convinced, in large part by Steve's arguments, that prices would be coming down.
I sold the unit for an 80% profit, and that was after a 8% reduction in price. So, I feel a certain gratitude to Steve who, in his own way, demonstrated to me the error of my arguments, and in the process, assisted me in turning a nice profit I am quite certain that the unit I recently sold has already gone down in value and will continue to do so for several years.
For those of us who are actual investors and players in this market, it is good that we have people who challenge our beliefs as to where the market will go. Steve's arguments have been consistent, persuasive, and so far correct. For that, I respect him.

But those points don't stick, do they?

Like zorter making amazing claims that the 12x-15x figure should be 20x, without having a clue what he was looking at. vverain claiming that housing wasn't leveraged. JuiceMan claiming that rents are rising even though people are losing their jobs. I forgot who saying I was quoting Al Roker when I was, in fact, quoting cnbc.com's personal finance editor being interviewed by Al Roker.

So you see, Tony, I don't make these figures up. They're real. Time people started to listen.

Thanks, mh23. No doubt that real estate has been a good "investment," and no doubt why it was when it typically isn't. My very first post here on or about December of last year was that it's far cheaper to rent than to buy, and that's an ominous signal for housing, since rentals and purchases are interchangeable goods: each with its advantages and disadvantages. But in the long-run, no one will pay a massive premium for marble countertops. Research has borne this out.

I also own, and sitting nice and comfortable in Fire Island reading the newspaper. I plan to do some major renovations in the off-season, including a new kitchen. Prices have come down somewhat, but I didn't buy to flip - I bought it as a place to escape Manhattan on the summer weekends. Owner-occupied housing is to be enjoyed and used, not viewed as something whose price goes up forever, because it doesn't: prices are permanently constrained by income and leverage.

I agree. Every time I made money in real estate it was when I was actually using the place, or when I inherited from my parents, who used it. Right now I am just starting to think about looking for a three bedroom ocean view condo in Miami. I have two kids and, G-d willing, I hope to have at least one more. My wife and I love Miami and, before the kids, we went down four or five times a year. Now that the kids are getting older, and so am I, I am thinking of buying something to own and maybe use for part of the year when I retire.
What are your thoughts, Steve, about the Miami market. I would ideally like to be on or around South Beach, but definitely within walking distance to Lincoln road. I think most of the land has been developed, so I would probably need to buy a resale in one of the newer buildings.

LTCM placed huge bets that spreads would return to historical norms and they lost big. Observing that price to rent multiples have risen from their historical norms does not tell you when they will return to normal, or even if they will return to historical norms, or if they won't return to historical norms, to what level they will return to.

I think the gist of Steve's argument is that it is irrational to pay so high a multiple, and that consequently demand will reduce as people choose instead to rent, thereby lowering the multiple. However, as is amply demonstrated on this board, there is short supply of rational thought in the Manhattan real estate market. As Keynes long ago pointed out, markets can stay irrational longer than you can stay solvent.

Alternatively, the argument is that demand will be reduced (due to costs of mortgages and the difficulty of getting them and the loss of income from the financial sector, etc) and supply will increasing (new construction, people who can't continue to afford their housing).

Their certainly are a lot of units on the market, but a supply curve maps price to quantity sold, and all these units have not sold so they could be thought of as contributing to the supply curve above the intersetion with demand ("unrealistic" sellers?). And indeed there is a ton of cash on the sidelines, but these buyers have not bought and so they can be thought of as contributing to the demand curve below the intersection with supply ("unrealistic" buyers?).

Don't get me wrong; I'm bearish Mannhatan housing, but I think the outcome is far from certain.

It would be interesting for StreatEasy to run a prediction market like HSX.com or PPX.popsci.com. Then we could all bet on binary options on price to rent multiples or futures on Miller Samuel's statistics. (You can trade NY housing futures on CME, but they are based on the NY Metro Case-Schiller index which is quite broadly based).

Steve once again, you pay 4200 per month rent for a 2 bedroom 2 bath apartment in Chelsea, so that apartmrnt is worth
4200x12=50,400x15=756,000. Do you honestly think that is what your apartmrnt is worth???Can you answer that question with a yes or a no answer?

jrd - That is the perfect argument. Nobel prize winners lost on the same bet that things always come back to the statistical norm. They were wrong too.

Guys, sorry, the LTCM example is incorrect. LTCM was right about ratios returning to historical norms. However, they lost (and lost huge), because they couldn't predict when/how it would happen. More importantly, their complex leveraged bets made it impossible for them to weather the pain while they got there. As an aside, they also lost big because they made a very wrong assumption that various markets were uncorrelated.

A simpler analogy is tech/internet stocks in the late 90's. Using Nasdaq for a proxy, there were many smart people who made the (correct) observation that it was way overvalued. However, if they shorted, say in late '99 at 3000-3500, they lost their shirts as it continued its run-up to 5000 and they couldn't meet margin calls. Then it crashed, eventually going well below 2000, and never touching 3000 again in the following 8 years. So the thesis of the guys that shorted in late '99 was dead-on, but timing killed them.

Very few people are smart enough to time any market perfectly (stocks, RE, commodities, etc.). If you're good enough to do it (like mh23 who bought on the way up, and sold at/near the peak), more powere to you. But Steve's fundamental observation is right. If buy-rent ratios are way above their historical norms, they will eventually come back down. No promises of when, of course.

zorter, the main figure Steve has been using on his blogs all over the place is 12x rent. So the example should be 4200x12=$50,400x12=$604,800. I would buy that apartment or anything similar for this price. Of course that is the point that Steve is trying to make but that price is ridiculously low even if you see another decline of 10-15%. And Steve still needs to answer your question with a 1 word answer. Can he do it?

steve, how do you calculate the 12-15x figure with maintenance fees, etc.? Is the 12-15x rule of thumb rent minus maintenance x 12 months x ~14 years? Or does maintenance not figure into the rule of thumb?

there is no "magic" about 12-15x. a home is an investment like anything else. you can earn a 4% yield in risk-free government bonds; a 5-8% yield in somewhat riskier bonds; and a 5-10% earnings yield on stocks (generally speaking). what yield you require on a house is up to your perception of risk and cash flow growth. i personally would argue that risk is high (correlated to banking income) and cash flow growth is flat or negative (falling rents) - therefore i would want a higher yield than all of the above (argues for the low end of 12-15x). the market is way out of whack due to the bubble - it will eventually return to normal but when that happens is anyone's guess (and could also be accomplished by rising rents in theory).

newbuyer 99: I find it hard to imagine that LTCM could be described as "right" in any pragmatic sense of the word. If you re-read my post, you will see it is all about when, if ever.

I agree that LTCM was not right in any pragmatic sense of the word. The issue I took with your post was with the "if" ratios ever return to historical norms. They do. The tough part is predicting when.

Like you, I am bearish on Manhattan real estate, and, I agree that the outcome (and especially the timing) is far from certain.

newbuyer99: why exactly are you sure that ratios will return to historical norms? Many things have changed. It used to be that people came to NYC, were successful (or not), got married, had kids and moved to Westchester or wherever. As can be seen by data from private schools and cencus data, this is not happening to the degree it was, which has a nontrivial impact on both supply and demand. I happen to own a 4BR coop unit. What are my alternatives to rent? They are quite few. So why exactly is the rent ratio meaningful for me? It isn't.

So I am willing to believe that the ratios for Studios, 1BR, and 2BR may well return to historical norms. But I think that for larger apartments, it will not. Which is not to say that that I know where it will go or when.

jrd, you are making an argument that supply/demand for housing may have changed structurally. that is certainly possible however if that were the case it would also manifest itself in higher rents. I do not believe that under normal circumstances people will pay 30-50% extra to own vs rent. That only seems justified if you expect price appreciation. And expecting price apprecition when rents are flat/down is essentially the definition of a bubble.

These ratios are like P/E ratios for stocks. Historically stock P/E's are around 10. What justifies a high P/E of stocks? If everyone only buys stocks with P/E of 10, noone will any make money. From my limited experience in stock investing, I have made the most money on stocks with high P/E's. Of course, I lost the most money on those as well. Timing is everything and Steve may be right, not because of the ratio, but because the trend is your friend.

Companies with high P/Es are often fantastic investments - as long as the have high cash flow growth to go along with it. What P/E would you put on a stock with declining earnings that is correlated with the banking cycle? You should put that same multiple on NYC apartments. Declinging cash flows (e.g. rental income) do not make for high multiples. This would be very different if it were 2003 and the market was beginning to look forward to several years of rental income growth.

steve has waffled more than Joe Rogers Sr. in regards to the 12x vs 15x ratio. He continues to post articles referencing 15x and then takes credit for brilliance. 95% of his posts on this board have been 12x except for the one where he said he would buy at 15x. How much sense does that make?

"you can earn a 4% yield in risk-free government bonds; a 5-8% yield in somewhat riskier bonds; and a 5-10% earnings yield on stocks (generally speaking). what yield you require on a house is up to your perception of risk and cash flow growth."

mbz, you are over complicating this. Owner occupied real estate is not an investment but rather, a place to live. If you really want to think of it as an investment, the only comparison that makes sense is comparing it to the cost of renting. Simply, as soon as you make $1 more from buying a place than renting something similar, you have made a good investment. 10%, 20%, and 80% returns on the home you live in are great, but not something you should count on. People need to stop thinking about their homes as an investment vehicle. That’s what got us into this mess. Buy a place that you want to live in, enjoy it, and stay in it long enough for it to be less expensive than renting over that same period. It is a simple as that.

Steve, you ended your first post with a question mark, but no question. Was there a question you wanted answered?

Ask steve why 15x is too much? Ask him why when I gave him practical numbers at 16x that made perfect sense to buy, he can't handle it and babbles on that 15x is the number without taking into account tax benefits? Which really makes 15x totally irrelevant. steve is a broken record of misleading information.

stevejhx I believe zorter asked a very good question. If you are paying 4250 per month in Chelsea for a 2 bedroom than according to your formula it's only worth 756,000. Is that correct?

this has turned into an interesting post for me given the 12x-15x rule and how it might apply to two threads that i'm currently soliciting advice on:

http://www.streeteasy.com/nyc/talk/discussion/4333-star-tower-in-lic-and-comps
http://www.streeteasy.com/nyc/talk/discussion/4334-salary-numbers-profile-need-advice

the unit in question is a 1BR in LIC that i can get for $400K. where i believe i can get rent of $2300/month. that would be 2300 X 12 X 15 = $414K. actually, if $2300 were indeed an accurate rent number, that would put the unit at about a 14.5X ratio. that makes me think that the unit is actually "priced well". comments?

jrd: You make a very good point on larger apartments and their scarcity. My wife and I would like to raise our kids in the city, and so are facing / will face the same issue as you. However, mbz answered your question better than I could have. If the market drivers have changed (more families wanting to stay in NYC), the market will adapt too, and management companies or owners will find ways to provide more larger apartments for rent.

As an aside, I wonder whether the trend of families staying in the city is a long-term change or a side-effect of the economic/wall street boom (i.e. more people being able to afford to). I've thought about this a bit, and don't know what the answer is. Will be interesting to watch in the next few years.

People won't advertise when they laugh all the way to the bank. I sat idle in the rental market in 2000 waiting to see what the market was doing - I could have bought a house for $140,000 but no like an idiot listened to every media gaff going on - at the end of 2003 the same property sold for $280,000 - and I had been paying rent all that time waiting for the market to fall. Now I have 15 properties because I made the choice not to sit on the fence anylonger - waiting, waiting, waiting, - for what - it does not work that way... The market will go again it's just a matter of time - if you have not bought - you should.

JuiceMan, go back to my earliest posts. I said 12x was the norm, but that I would consider dipping my toe in at 15x.

Go back to my other threads - much of whether it's 10x or 15x depends on interest rates.

I've always been consistent.

To the others, I'm glad I was missed from my few days of not posting. The weather was too nice.

Zorter, first I pay $4500 not $4200, and yes, I would consider buying my apartment anywhere between 12x and 15x rent. Unfortunately right now a very similar building across the street sells for 24x my rent, which I would NOT pay. Nor would I pay 20x or 18x or 16x.

jrd, to compare real estate to Long-Term Capital Management is ridiculous. LTCM a) was trading securities, not capitalizing rent; and b) was in a highly liquid market, unlike real estate. I can't even comment on the analogy.

The ratios MUST return to the long-term mean because they always do, even in Manhattan. The feeding frenzy is over, inventories are rising, incomes are falling, investment banks are deleveraging (and hence, their incomes will fall). Who is left with the money to buy.

LICC, you're difficult to abide. You gave a 16x multiple that "proved" your point when, in fact, it did nothing of the sort. Your multiple included the "tax benefit," which is NOT included in the price-to-rent ratio, which is a naked ratio of price to, well, rent.

If you want to use the tax benefit / opportunity cost, then use the imputed rent model. Otherwise, stop typing and spewing nonsense.

csn, "that price is ridiculously low." No it's not. It's the price of just a few years ago. Owner-occupied residential real estate does not and cannot increase at those rates - better than Berkshire Hathaway on a return on investment basis - because it is physically constrained by incomes and leverage. There are no economies of scale to take advantage of in owner-occupied real estate. There is no increase in productivity. It is merely an EXPENSE, albeit one, like any prepaid expense, that is offset by modest earnings, be they increase in value or interest.

You're seduced by current prices and therefore believe that they are rational. I remember when gas cost 27 cents a gallon. We will never return to that price level. I paid $4.69 a gallon on Thursday for 93-octane. Am I used to that price? No. Do I think it will fall? In the short-term perhaps, in the long-term no. Why? Because unlike owner-occupied real estate (or gold, for that matter) it is an INPUT. Owner-occupied real estate is not an input.

inoeverything, I sold an apartment on South Beach for $1 million 2.5 years ago. An identical unit now is now listed for $675,000. The buyer of my apartment had the intention of buying it and flipping it in 2 years. Good luck.

mh23, I'm biased when it comes to Miami: I hate it. But if you move, be very certain where you want to move to and whom you want to move with. Lincoln Road is great for a day or a week; on a more permanent basis it's very small, and very boring, and south of Washington very sleazy and not getting better. Ditto Collins.

You might look at Las Olas Boulevard in Ft. Lauderdale.

"south of Washington"

Tee-hee-hee. I have Long Island in my mind. "East of Washington," is what I meant!

I guess not.

okay so you answered the question you feel your Chelsea 2 bedroom apt. in not worth a penny over 810,000.

That is incorrect spunky, steve is saying he wouldn't pay a penny over $648,000 (12x) for his 2 bed apt in Chelsea.

ccd, you "guess not" what?

spunky I answered that I would not pay over 15x the rent to buy the same place, 12x is the historic mean, sometimes it goes lower. The ultimate price would depend on interest rates and a few other factors peculiar to an individual unit (which I've always maintained had to be taken into account) such as property taxes (lower for co-ops than condos) and maintenance / common charges, and which direction I thought the market is going, but the short answer is yes, $810k.

And that's not so unusual since that's what they were going for in 2003-2004, and I've always maintained that barring divine intervention, that's where we were headed back to. The 100%+ increases we have seen in the past 5 years are unsustainable - not even Berkshire Hathaway makes that much money so consistently.

30-40% corrections throughout NYC by this time next year. That will bring us back to the 12-15x rent ratio. I would be embarrassed to admit buying in the environment. I'm still shaking my head.

so steve, is your baseline 15x or 12x? Which is it?

Oh, JuiceMan, there you go again! I've always been consistent: historical norm is 12x, I might be willing to dip my toe in at up to 15x depending on exogenous factors such as the ones I listed.

Somewhere between $648,000 and $810,000 is about right, depending on the specifics of the unit.

"30-40% corrections throughout NYC by this time next year."

Care to place a wager dco? I'll bet you $10,000 that median prices in Manhattan will not fall by 35% (I picked your midpoint) by this time next year. We will use streeteasy's sales search engine with the "must have address box" checked. Today the median price is per sqft in Manhattan is $1,178. On 7/21/2009 if the median price per sqft is $766 or below, I will pay you $10,000. If not, you pay me $2,500. How about it?

JuiceMan- Be realistic. We both know we are not going to wager money. You know that you also can't go off median prices, they are not reliable enough. I'll be the first to tell everyone I was wrong. The question is will you.

I guess AMEX will be the next to give out pink slips.

"We both know we are not going to wager money."

Says the guy with no money.

steve, I guess there was no question.

Surely if you are confident enough to make 30-40% predictions you should be able to put something behind it?

ok dco, can we make it a little interesting? How about this, we use streeteasy median prices as of 7/21/08 and 7/21/09, "must have address box" checked, and a 30% reduction off median prices ($825 a sqft). The winner picks a charity and the loser donates $100 to it. Anyone on this board can play. Just respond to this thread with either under $825 psft or over $825 psft median target and the dollar amount you are pledging. Next year, those who responded correctly will pick the charity and those who chose incorrectly will donate their pledged amount to it.

Now dco, surely you can come up with $100 for a worthwhile cause and some bragging rights?

"I guess there was no question."

Question for what?

JuiceMan, dco is correct: you can't go off median or mean prices because they are skewed for which apartments sell in a quarter, not how much any individual apartment sells for, or what apartments are for sale at what price. Therefore you need to go off the price of a single apartment. For that there will not be enough information in 1 year since on average fewer than 10,000 apartments change hands in Manhattan each year, so there will be very few of the same units selling more than once.

You could try the price-to-rent ratio, but then you'd get into an argument over what is an "equivalent" unit, which I know you vehemently disagree with me on.

steve keeps sticking with this 12x-15x analysis even though when applying it in real terms it completely fails. Let's try again. Let's say someone takes an $800k mortgage at 7%. The monthly payment would be $5300-$5500. Say another $1500-$1700 for common charges and taxes. Your total after tax monthly costs would be between $5,000-$5,500. A $5,000 rent at 17x is just over $1 million. A $5,500 rent at 17x is $1.12 million. If those of the types of rents you are facing and you can buy at those prices, that seems like a pretty good deal. Especially considering that over time your rents will just keep rising and rising.
It's tough when practical numbers make your arguments look silly steve and dco.

Juiceman, count me in. I'll take the over.

Leave it to steve to say that looking at one individual apartment is a better indicator than looking at the mean or average of the whole market. The man doesn't know the meaning of the word objective.

stevejhx for president!!!

"you can't go off median or mean prices because they are skewed for which apartments sell in a quarter, not how much any individual apartment sells for, or what apartments are for sale at what price"

Now you are going to lecture me on the deficiencies of using median as a statistical measure for real estate. You are a piece of work. Median is as close as we are going to get and it could "skew" either way. Stop wasting my time with your nonsense.

streeteasy doesn't have median on sales, just list prices. The bet is 30% off the median list price unless someone can point to another way to measure it (sales would be preferable if we have the data).

No matter what the number, I'm quite certain no one will agree. I still stick to my prediction. I can hear it now, it's not 30, that's 29.99999. I'm actually laughing typing this because I can actually see the words a year from now.

JM- Why don't you put your Street easy rep. on the line and give your prediction for 7/21/09. While were at it, lets hear from the "crew".

um, not to get technical here, but medians really aren't considered susceptible to skewing. it is means that are skewed by outliers. medians typically fix that skewness problem.

i would like to act as escrow for these $100 bets going on.

dco this isn't that complicated. If the number is $826 or higher, you lose. If $824 or lower, you win. We'll even let you win if it is $825, how is that?

dty, you are correct, but now steve will give you some incomprehensible reason why you are wrong and he will cite to some article that has nothing to do with this issue to back his argument.

i find this conversation about price-to-rent ratios fascinating because when it comes to real estate, well, i'm dumber than you. here's a quick report on the top 3 links when you google the terms "price to rent ratio housing" (google the terms without using quotes).

link #1: http://bigpicture.typepad.com/comments/2008/05/housing-price-t.html
suggests that, nationally, the mean P-R ratio is closer to 21.5 and has never really been below 19. however, these are national numbers, and still, the tenor of Steve's arguments are certainly consistent with the chart.

link #2: http://money.cnn.com/magazines/fortune/price_rent_ratios/
reports that the P-R for new york (city, i assume) as of june 2007 was 17.8, while the 15 year average was 11.7, suggesting a 35% correction is in order to reach the 15-year average. but these numbers seem suspiciously LOW, don't they? (perhaps because the 4 other boroughs are included?) the national average of 22.8 is closer to the approximate 26-27 figure reported in the same time period under the chart in link #1.

link #3: http://calculatedrisk.blogspot.com/2008/07/price-to-rent-ratio-update.html
shows similar data, normalizing the ratio to 1.0 at 1982, showing that the ratio is about 1.5 right now, again suggesting that about a 33.3% correction is due.

p.s. please do not post a reply to this thread by calling me names or being sarcastic. i have no desire to participate in the pissing-contest aspect of this post. i'm just reporting here, folks...

the last link i'd like to mention is this one:
http://www.thestalwart.com/the_stalwart/2006/05/the_house_price.html
it is the last link on the google search page and is therefore subjectively chosen by me (unlike the 3 links above). here's an excerpt that i found interesting:

"To bring the ratio of prices to rents back to some sort of fair value, either rents must rise sharply or prices must fall. After many previous house-price booms most of the adjustment came through inflation pushing up rents and incomes, while home prices stayed broadly flat. But today, with inflation much lower, a similar process would take years. For example, if rents rise by an annual 2.5%, house prices would need to remain flat for 12 years to bring America's ratio of house prices to rents back to its long-term norm. Elsewhere it would take even longer. It seems more likely, then, that prices will fall. ...But even if prices in America do dip, insist the optimists, they will quickly resume their rising trend, because real house prices always rise strongly in the long term. Robert Shiller, a Yale economist, who has just updated his book “Irrational Exuberance” (first published on the eve of the stockmarket collapse in 2000), disagrees. He estimates that house prices in America rose by an annual average of only 0.4% in real terms between 1890 and 2004." [dated may 2006]

"p.s. please do not post a reply to this thread by calling me names or being sarcastic. i have no desire to participate in the pissing-contest aspect of this post. i'm just reporting here, folks..."

Oh come on now... it's phun! Currently, I'm the Wile E Coyote according Goldie Oldie. =D

last post. what strikes me about the 3 links i mention above is the variance in the P-R ratios reported. the 15-year average for new york is 11.7? as of last year, it is only 17.8? the national average is 21.5? what it makes me think is that if indeed P-R ratios are a correct way to view the world, then maybe using a 12x-15x bright line rule is not the way to go. that is, there are clearly differences in the way that these numbers are being calculated, and insofar as i don't know the adjustments that are made, i'm gonna have to assume that a straight-up simple "price divided by rent" calculation and then applying it to a 12-15x rule may be a bit too restrictive. that is, maybe coming in at 17x is not a death sentence, since there seems to be some wiggle room in calculations.

"what it makes me think is that if indeed P-R ratios are a correct way to view the world, then maybe using a 12x-15x bright line rule is not the way to go. that is, there are clearly differences in the way that these numbers are being calculated"

Really? You don’t say? steve blabs incessantly about 12-15x but has no idea 1) how these numbers compare with Manhattan (he often argues that Case-Shiller isn’t applicable to Manhattan however, these numbers are built in a similar way, yet THESE are law) and 2) he has no clue where Manhattan is today (he continues his 24x rant which is even dumber). It’s hysterical, he has a little following on this site and he is going to lead them right into the gutter.

Here is a guy that states that median's are subjective and cannot be used as an accurate measure in real estate yet bases all of his logic and "analysis" on 12-15x and some economic white paper that uses a tragically flawed formula (which for all you newbie’s out there looking for substantiated rebuttals -I read, analyzed, and challenged steve on. His response was that “all formula’s are flawed in one way or the other”) Why is it that the most utilized, recognized, and accepted measures are flawed, yet any subjective, bullshit formula that steve can find an article on is treated as law? Tell me steve, why is your 12-15x “analysis" any different than using Case-Shiller as a barometer for Manhattan?

Anyway, dco so that means you are not taking the bet? Why am I not surprised? You and steve are so confident in a 30-40% correction yet you can’t bet $100? Why should anyone listen to anything you have to say? Why not just take the anon3 route and say NYC will be down 80%?

I like the idea of a bet.

The issue is how to create a mutually agreeable framework. There are unquestionably difficulties with means and medians in keeping the analysis apples-to-apples. I have a suggestion. It's a little bit of work, but it helps address the comparability question and should minimize "product mix" issues.

Why not come up with a selection of buildings that would comprise your "index", each of which building has sufficient # of units and transaction volume to serve up meaningful y-o-y statistics. These would therefore need to be large coop/condo buildings.

To make the comparison, you simply calculate the $/sf average for each building's closed transactions say over the last three months. Average all of the buildings' average $/sf to come up with your 2008 starting point. One year from now, do the calculations again based on closings in the trailing three months 2009. There's your comparison.

To avoid selection bias, the bulls (i.e., Juiceman) can pick out 15-20 buildings. The bears (i.e. Steve) can pick out 15-20 buildings. That should balance out any gaming in the analysis. You could even take it one step further and say that each side can disallow maybe 3-5 of the other side's selected buildings.

It's not difficult to do. Could be very interesting. And I'm definitely in.

What an interesting discussion!

When I first moved to NYC in 1989, the local re market was in the "bust" part of a boom-and-bust cycle, and I recall reading an article in New York mag that quipped, in NYC, real estate is a "spectator sport."

It was such a good line that I never forgot it. It has never stopped being true.

I tend to be bearish on the economy, and on NYC re in particular. But I really think the question you are all dancing around is, "how bad is this down cycle going to be for the economy"?

This is white collar recession, a Wall Street recession. A good chunk of the Street has already laid off, and yet another equal-sized chunk is going to be laid off over the next 12 months, as the financial shops face reality, and realize that many of their former business lines are kaput.

Wall Street is the great engine of the city. And it isn't just hedge managers buying the penthouse apartments, it's the assistants and the junior analysts and the out-sourced contractors. Without this industry, the city is going to recoil into a rather bad case of traumatic shock.

And real estate is going to take a real thrashing. Maybe certain addresses will hold onto much of their value; maybe certain types of structures, like townhouses, which are not being made any more, will too.

But by and large, the market is going to tumble.

faustus, I like your idea because it makes sense. However, I have learned in the past that trying to discuss anything that makes sense with steve will derail into hours of endless babble about meaningless details that have nothing to do with anything. Additionally, the index will ultimately end up with the top 15-20 buildings in the West Village and the bottom 15-20 buildings in Harlem, which would be meaningless.

Juice - so make it a requirement that each group pick 20 bldgs - 5 below 14th street, 5 between 14th/59th, 5 between 59th/96th, 5 above 96th (or some variation thereof). Not difficult. It should all (almost) wash anyway if each group picks half the buildings.

I think you are on to something faustus.

Steve - thoughts?

I like the faustus bet. I'm in for the under. I don't think I'll win but I like the concept! $200. And I vote for the "under" charity to be the Leukemia and Lymphoma Society. I'll leave it to you guys to pick the buildings. Let's make it 25% to make it more fair? If not no sweat I'm in either way.

LICC: "steve keeps sticking with this 12x-15x analysis even though when applying it in real terms it completely fails."

You keep on singing the same song, and you keep on being wrong. 12x annual rent is the ratio between the rent and the purchase price of a home. WHERE IN THAT RATIO DO YOU SEE TAX BENEFIT?

Nowhere. Because it's not there. It is in the IMPUTED RENT model, and feel free to use it.

12x is the historical mean. That mean is affected by (primarily) interest rates; when they are low, the ratio can go higher, and vice versa. THE TAX BENEFIT IS NOT COUNTED IN THIS METHOD.

That said, there are factors unique to NYC that are not included in the price-to-rent-ratio, most especially differences in common charges and taxes. NYC has the world's most convoluted property tax system where properties worth $1 million can be taxed anywhere between $5,000 and $20,000 a year. That variance needs to be taken into account as it is material, the same with common charges given places like Battery Park City and other land-lease buildings, tax abatements, etc., which are also material.

But the tax benefit is NOT counted. Nor is the opportunity cost of not investing in another asset class.

Case closed. Go back to high school.

"Now you are going to lecture me on the deficiencies of using median as a statistical measure for real estate. You are a piece of work. Median is as close as we are going to get and it could "skew" either way. Stop wasting my time with your nonsense."

JuiceMan, you're getting to be as sad as LICC: mean and median are significantly skewed each quarter depending on the universe of apartments sold. 15 CPW and the Plaza SIGNIFICANTLY skewed last quarter's mean and median; following the cost of a single apartment in each of those units does NOT skew the sample.

Only a fool would offer such a bet, or take it.

"medians really aren't considered susceptible to skewing. it is means that are skewed by outliers. medians typically fix that skewness problem."

That's entirely incorrect. Medians are indeed susceptible to skewing. Say 5 apartments sold in Manhattan this quarter. 4 of them sold for $40 million at 15 CPW. 1 sold for $199,000 in Inwood. The mean is $32,039,800.00, the median is $40 million. That is a 20% difference in favor of the median.

Now let's reverse that. Say the next quarter 4 apartments sold for $199,000 in Inwood, and 1 for $40 million at 15 CPW. The mean is now $8,159,200.00, the median is $199,000, the difference between the two is now 4000%.

Yet, between the two quarters, the mean and median are EXACTLY the same: $20,099,500.00. That's why this "bet" is so absurd: it all depends on which apartments were sold when.

dumberthanyou, the reason you have those different numbers is because they are calculating different things. The first chart is not explained in the full version of the materials:

http://www.frbsf.org/news/speeches/2008/charts.pdf

but I believe that it is the imputed rent, and that is why the figure is higher - it includes the tax benefit, the opportunity cost, the risk premiums, expected increase / decrease in sale prices, etc. That is, "Third, differences in expected appreciation rates and taxes can lead to considerable variability in the price-to-rent ratio across markets," from:

www.ny.frb.org/research/staff_reports/sr218.pdf

indicates that it is being calculated using the imputed rent model, which will give you a higher result.

The more typical price-to-rent ratio is the second of your posts, as calculated by Shiller and Fortune Magazine, among others. The third is the deviation from the norm.

LICC: "Leave it to steve to say that looking at one individual apartment is a better indicator than looking at the mean or average of the whole market. The man doesn't know the meaning of the word objective."

Standard and Poor's: "The S&P/Case-Shiller Metro Area Home Price Indices use the “repeat sales method”
of index calculation – an approach that is widely recognized as the premier methodology for indexing housing prices – which uses data on properties that have sold at least twice, in order to capture the true appreciated value of each specific sales unit."

www2.standardandpoors.com/spf/pdf/index/SPCS_MetroArea_HomePrices_Methodology.pdf

JuiceMan: "Here is a guy that states that median's are subjective."

Not ever did I say that. I said they are skewed based on a small set of data.

"he often argues that Case-Shiller isn’t applicable to Manhattan"

I said the raw data from Case-Shiller exclude 99.1% of all Manhattan data. I have ALWAYS said that their methodology is FULLY applicable to Manhattan.

"some economic white paper that uses a tragically flawed formula"

Yes, here is JUICEMAN, with all his economic expertise, criticizing the methodology used by the Federal Reserve Bank of New York, yet he NEVER provides his own model or his critique of what's wrong with it. In fact, if I dig deep enough, I can find the quote where he states that he didn't even understand the model.

Pick a different one then. But an ACCEPTED one.

"he continues his 24x rant which is even dumber"

I believe you agreed with this on another thread.

http://www.streeteasy.com/nyc/talk/discussion/4209-rent-ratio-heat-maps

All prime areas of Manhattan are 24x. Areas with lesser access to transportation are 14x to 16x. The slums are cheap.

"Why is it that the most utilized, recognized, and accepted measures are flawed, yet any subjective, bullshit formula that steve can find an article on is treated as law?"

What are that "most utilized, recognized, and accepted measures"?

And all formulas are flawed.

"which for all you newbie’s out there looking for substantiated rebuttals -I read, analyzed, and challenged steve on"

No you didn't. You claimed that there was some flaw or another, you claim you wrote the authors and they didn't bother to answer you. Because they are tenured Columbia Economics professors and Fed staff members, and you are, well, the JuiceMan.

"Tell me steve, why is your 12-15x “analysis" any different than using Case-Shiller as a barometer for Manhattan?"

Case-Shiller's data can't be used for Manhattan; their methodology can, and yet again I invite you to go here:

http://350bleecker.com/policy/sales.html

and compare what happened to the prices of identical or virtually identical apartments over time.

"why is your 12-15x “analysis" any different"

Because they're completely unrelated, which someone as "smart" as you should know: price-to-rent compares purchase prices with rental prices. Case-Shiller compares the sale price of a single unit of housing over time. How could you possibly equate the two?

faustus: "To make the comparison, you simply calculate the $/sf average for each building's closed transactions say over the last three months. Average all of the buildings' average $/sf to come up with your 2008 starting point. One year from now, do the calculations again based on closings in the trailing three months 2009. There's your comparison."

Because you are still comparing apples to oranges, studios to penthouses. The result - since the sample is so small - will likely be highly skewed as the methodology is highly flawed.

No wonder the JuiceMan likes it so much.

Steve - come on, let's think about this.

For the analysis to turn into a comparison of "studios to penthouses", as you say, would require that (1) in 2008, the sales in each building be heavily skewed towards one specific size apartment, (2) in 2009, the sales in each building be heavily skewed toward a much larger/smaller size apartment than in 2008, (3) this difference be consistent across most of the buildings in our index.

To address your concern, (1) I would doubt very much that in 2008 the apartments sold in the universe of buildings we choose is so heavily skewed towards one specific size of apartment. Bear in mind we're looking at 30-40 of the larger apartment buildings in Manhattan, each of which should have a good sample size of transactions.

(2) Whether or not there's size-bias in the 2008 sample, who can say whether there will be size-bias in 2009 (across ALL the buildings in our index)?

(3) Steve, even in the unlikely event that buying patterns change SO MUCH from one size apartment to another between 2008 and 2009 to have any measurable effect on the outcome of the bet (very doubtful), how do you know which direction it will go? Why do you automatically assume that it would favor Juiceman when it could just as likely favor you?

Look, I think your concern is pretty minor and that variations will wash unless you really, truly, think that in one year buying patterns will suddenly shift dramatically so that everyone's buying 3BRs and nobody's buying studios/1BRs. However, if you're so concerned about this then let's narrow the analysis so we only look at 2BRs and 3BRs. But if that's the case, each group will need to pick a larger # of buildings so we have sufficient data points.

ok, this is totally a technical point, but Steve, your point about medians is quite disingenuous. you bring up an example where you have a sample size of n=5 apartment units. it's easy to cook up anything you like with such a small sample. Stats 101 says the law of large numbers comes into play at about 30. and for our purposes, it is quite reasonable to say that the sample size is in the thousands. in which case, the median is actually a fairly decent indicator of some sense of "location" as far as the distribution is concerned. ... certainly less susceptible to skewing than the average. long story short: the median isn't a bad place to start when thinking about judging a sample. (though certainly not the end all.)

i don't know about the other points you make about the links i included -- because when it comes to real estate, i absolutely guarantee you that i am dumber than you. however, you do not corner the market on logic. so i stand my case on the median being a preferable measure to the mean if you are wary of outliers skewing interpretation.

btw, my purpose was not to shoot you down in any way, but to provide some (hopefully) useful links to interested readers. actually, i thought i helped your cause overall.

i can agree with you that the S&P/Case-Shiller Metro Area Home Price Indices use of the “repeat sales method” is very likely a superior methodology because it, in essence, controls for what the stats guys call cross-sectional variation (since you're using each home as its own control). it is, however, still susceptible to time-varying trends. and there are also issues of "endogeneity" and selection bias since these are homes that did indeed have repeat sales -- and the very reason for why these units had repeat sales is probably not sufficiently controlled for. but hey, no method is perfect. and my only point is that at least this method is logically consistent.

but yeah, medians are cool.

steve, I hope you are happy in your ignorance, because you are quite ignorant. First, your mean/median discussion is so flawed a 5th-grader wouldn't have made the mistake you did. Let me explain this to you. The median would be the middle number in an ordered set of data, or the average of the two middle numbers. As dty said, by using such an unrealistic and small sample size, you dishonestly manipulated your example. Just add one more sale to your example and it changes completely. This is typical of you, using unrealistic examples to back your small-minded arguments. In a sample size we are talking about (streeteasy listings), the median would not have this big skewing effect that you claim.

As for your idiotic application of the rent-ratio analysis, you admit that the tax benefit is not counted. Then why apply it to a practical discussion on real estate values! We do not live in a no-tax society. You are using an analysis that does not count the tax benefit and making arrogant conclusions about real-life real estate. You have yet been able to dispute my actual numbers that show your analysis to be completely wrong. More disingenuous commentary by steve. Do you not mind that everyone here sees through your crap? I guess it is hard to teach an old dog new tricks.

Hey, its possible that the trend skews toward smaller apartments as well. Its a BET, there's an element of risk here. But its going to be for charity so let's take it. The real advantage lies in the over because nothing has to happen for them to win. Real estate actually has to move down and not up or stay the same for the bear bet to win. Nevertheless, if we feel strongly about it let's take a chance people! Let's just do it at a reasonable level. I say 20%. That's a big enough move for a year to lay claim to bragging rights, and attract some bets on the other side. And if it happens to be 40% then dco can jump up and down and scream about how right his "analysis" was.

Besides all that, prices continue to rise or are flat at worst overall, and inventory is dropping.

Prices may trend downward slightly if the overall macro-economy continues to tank, but with 1.6 million people and 7400 listed units I don't think their going to come down by much. Also, given that interest rates are going to slide upwards, there may not be any great benefit to waiting much longer.

faustus, if you can correct for price variances in differently configured units, then you would have a far more accurate sample. Thus, say in a 30-story building, you compare the prices for the A-line to the A-line, B-line to the B-line, C-line to the C-line, etc., assuming they're all configured the same, then correct for the per-floor premium for higher floors.

It's still not perfect (nothing is) but it corrects for a huge bias.

"Why do you automatically assume that it would favor Juiceman when it could just as likely favor you?"

I don't. I assume that it COULD favor either one of us.

"you bring up an example where you have a sample size of n=5 apartment units."

It's not disingenuous; it was given to prove the point that certain luxury apartments can have on the mean and median, as happened in 1Q2008 with 15 CPW and the Plaza. It is true that this effect is diminished in large samples, BUT ... these new construction / conversions tend all to close in the same quarter, thereby having a material effect on the sample.

So, the argument goes, you would need to correct for that in any analysis you perform.

"[Cas-Shiller] is, however, still susceptible to time-varying trends. and there are also issues of "endogeneity" and selection bias since these are homes that did indeed have repeat sales -- and the very reason for why these units had repeat sales is probably not sufficiently controlled for."

No question on the micro level, but I believe there is sufficient correction for it by using large metropolitan areas each with different micro-markets, and each time series uses different datasets. That is, Houses A and B may be counted in Quarter 1, and Houses C and D are counted in Quarter 2, and Houses A and C are counted in Quarter 3.

"but hey, no method is perfect."

JuiceMan doesn't think so. In fact, he ridiculed me when I said that no model is perfect.

LICC: "The median would be the middle number in an ordered set of data, or the average of the two middle numbers."

Plug the numbers into Excel and get back to me.

"In a sample size we are talking about (streeteasy listings), the median would not have this big skewing effect that you claim."

You would need to prove that and I don't think you can.

"to back your small-minded arguments."

Wow! Come up with a generally-accepted method for doing what you want, then reconsider my "small-minded arguments." Like Case-Shiller, which you dismiss, but is widely regarded as the best indicator there is.

"As for your idiotic application of the rent-ratio analysis, you admit that the tax benefit is not counted. Then why apply it to a practical discussion on real estate values!"

Oh LICC, how I tire! I said that there are methodologies that you can use that include the tax benefit, there are others that do not. You can compare the results from each - valid - but you can't mix and match to achieve your end. I have ALWAYS advocated using:

1) Price-to-rent (includes just those two factors)
2) Imputed rent (owner's carrying cost - includes tax benefit / opportunity cost, risk premium)
3) Case-Shiller (how much short-term increase decrease has there been in a virtually identical unit)

I think those three models - I exclude price-to-income b/c I don't have enough data - will give you a very good picture not only of the market as a whole (macro), but of the value of a particular unit. But the Manhattan real-estate market is also skewed by an odd property taxation system (not income taxes, which is what you always talk about) that needs to be taken into account in any individual decision, as it can have a material effect on owner's carry costs. That seems reasonable to me.

"We do not live in a no-tax society."

Who said we did?

"You are using an analysis that does not count the tax benefit"

Not true. What I won't do is let you count a tax benefit in a model which does not count the tax benefit, nor will I let you count the tax benefit without applying a risk premium on ownership and the opportunity cost of not investing in another asset class, both of which you do not wish to include but are directly related to the tax benefit.

"and [are] making arrogant conclusions about real-life real estate."

"Arrogant conclusions"? Dude, get a grip!

"You have yet been able to dispute my actual numbers that show your analysis to be completely wrong."

Yes I have. Thousands of times. If you want to include the tax benefit, use a model and ratios that include the tax benefit. Don't take a ratio that excludes the tax benefit, then add the tax benefit back in because it supports your conclusion. That is GIGO.

"More disingenuous commentary by steve. Do you not mind that everyone here sees through your crap? I guess it is hard to teach an old dog new tricks."

Yup.

"with 1.6 million people and 7400 listed units I don't think their going to come down by much."

Inventory is 8,000 - 9,000. Historically that is a 1-year supply of apartments. The median income in New York County is $45,290; the median apartment price listed is $1,250,000. How many median income people are going to afford a median-priced apartment at those levels?

"Also, given that interest rates are going to slide upwards, there may not be any great benefit to waiting much longer."

If that is so, then prices will come down.

Ah, JuiceMan, from today's NY Times editorial on poverty: "No formula would be perfect."

JuiceMan: "His response was that “all formula’s are flawed in one way or the other”

LMAO.

Now, come up with yours.

faustus, see my point? This is why engaging steve is a waste of time. The guy thinks that median is an inaccurate measurement for a sample size of 10,000 yet is willing to throw around formulas from “Columbia Economics professors and Fed staff members” that were built for $200k houses in Albany as law. I don't have time to respond to all of his spin, conjecture, and crap posted above, and will not. The beauty of this board is we don't need steve for the bet. He can just sit on the sidelines and cry about the injustices of medians (HAHAHAHAHA).

Anyway, let's engage some level headed people and agree on a bet. I'll go back to the easy one, prices. The median psft today for all of Manhattan is $1,178. Again, these are prices, not sales. This should give the bears even more of an advantage because of all the "unrealistic sellers" out there that will have to chop their prices in a hurry. 30% off would be $825 psft on 7/22/09. It isn’t perfect, but it is equally deficient on both sides of the argument. Anyone besides steve have a comment?

http://www.bloomberg.com/apps/news?pid=20601039&refer=columnist_berry&sid=aFYsqGVoF5mI

Looks like 2Q numbers may not be that bad for the overall economy.

"The guy thinks that median is an inaccurate measurement for a sample size of 10,000."

Untrue. But moreover, go to streeteasy.com, scroll down to where it says:

"Medians for listings from past 60 days from StreetEasy data. Excludes some extraordinary properties. No representation is made as to the accuracy of this data."

Oh! Look at that! "Excludes some extraordinary properties."

What did I say?

And better: "No representation is made as to the accuracy of this data."

So you're going to place a bet on data that may or may not be accurate?

"let's engage some level headed people and agree on a bet."

Level-headed = bet on potentially inaccurate data. Good!

"yet is willing to throw around formulas from “Columbia Economics professors and Fed staff members” that were built for $200k houses in Albany as law.

Even if the Albany comment were relevant (not), you have yet to provide a model to measure current house prices.

"He can just sit on the sidelines and cry about the injustices of medians (HAHAHAHAHA)."

Let me rephrase this: "He can just sit on the sidelines and cry about the injustices of EVER INCREASING MANHATTAN PROPERTY PRICES (HAHAHAHAHA)."

Boy I've heard that one before!

JuiceMan: "I don't have time to respond to all of his spin, conjecture, and crap posted above, and will not."

Because you cannot.

"you have yet to provide a model to measure current house prices."

How can I provide a model or data when everything is inaccurate? What do you think is more accurate steve, the median psft on streeteasy's sales page or your blanket statement that everything is Manhattan is 2x more expensive to buy then to rent? Every post you make is inaccurate, yet you point out "extraordinary properties"?

"Because you cannot."

I can, have, but will not, because you are a waste of time and are the only one on this board that is incapable of having an intelligent, unbiased, thought provoking conversation. Post away steve, I don't care, this board is much better when you are ignored.

Go away, we are trying to have fun and treat each other in a civilized manner. Please take your douchebaggery somewhere else.

seriously, Steve, you need to stop harping about medians. it really does expose you.

"How can I provide a model or data when everything is inaccurate?"

That seems to mean that everything you post is subject to criticism, which you reject. But everything is subject to criticism, no model is perfect, but I've yet to see you tell me how you would evaluate property prices as a whole. You're going on median prices, which have some relevancy which I accept, but it is not something that you can place a bet on because they are influenced by the mix of apartments for sale at any particular time.

"I can, have, but will not."

"Veni, vidi, vinci," I think is what you mean. You don't have to do it again, just point out the post.

To quote you:

"he continues his 24x rant which is even dumber"; "your blanket statement that everything is Manhattan is 2x more expensive to buy then to rent?"

To quote me:

"All prime areas of Manhattan are 24x. Areas with lesser access to transportation are 14x to 16x. The slums are cheap."

"Why is it that the most utilized, recognized, and accepted measures are flawed, yet any subjective, bullshit formula that steve can find an article on is treated as law?"

What are that "most utilized, recognized, and accepted measures"?

You've yet to answer, yet I'm the one who is "the only one on this board that is incapable of having an intelligent, unbiased, thought provoking conversation."

"douchebaggery."

:)

DTY, I would ask you to pause and think a little bit more about the median sales price. I thought Steve was wrong at first also. He's not, but it takes a minute to figure out why. The median *value* of a unit in Manhattan is an extremely reliable number, but there is no easy way to determine what that value is. The sample size of trades does matter, as you said, however it is not randomly chosen, which creates a bias. If the sales in a period were random, it would be a pretty accurate number. And in a stable market, it is reliable because the unit mix period to period is stable.

However, in a volatile market like today's, it is not as reliable, because there is a systematic bias in the data. Specifically, people who no longer have access to 90%+ financing are being eliminated from the market. Buyers with more cash are probably more likely to buy more expensive units (no one is buying a $10m unit at 15 CPW with a 90% mortgage). As a result, the trades, and therefore the median trade, is skewed to the high end.

TenthStreet, the median gives a measure at an exact point in time. It is not forward-looking.

Where is steve getting this 24x number? By looking at one apartment in his building and generalizing the whole market? Didn't someone put a heatmap link up some time back that showed most of Manhattan at 18x or less?

DTY, steve posts lots of inaccurate, misleading nonsense that exposes him.

LIC, I agree, it measures a point in time. I don't think my comment contradicts that. I'm just saying that the median can be systematically biased because it is not random.

Keep in mind TenthStreet, we are talking about median price per square ft and not median price, so there is less of a bias. There is some however, because larger units tend to have a higher price per square foot – but we are really splitting hairs now. We would need a very large shift in the purchase of studios and 1 beds for the MEDIAN price per square foot to be impacted (not price…..PRICE PER SQUARE FT) Further, the predictions have been for prices to go down 30-40% by next year. Unless I missed it, I have not heard that prices will come down on studios and 1 bedrooms by 30-40% by next year, or prices will come down in new developments by 30-40%. What we get on this board is it is "2x more expensive to buy than to rent" and "prices will correct by 30-40% by next year". If the geniuses making these predictions would like to be more granular in their predictions, I welcome it. It seems that the only time they want to be more specific is when they are challenged on their idiotic predictions. Go figure.

It also seems that by steve arguing about the deficiencies of median as a statistic in measuring real estate, he is also admitting that he has no consistency in his claims. If I was bold enough to say it is 2x more expensive to buy than rent in Manhattan, the mix wouldn’t matter (certainly not on price per square ft). Same for the 30-40% prediction.

10th, i actually did allude to this up top. i don't like quoting myself so, i'll just paraphrase:
there are indeed issues of "endogeneity" and selection bias in the sample since the homes that sold may have sold for reasons that make them, by their nature, different from ones that didn't sell. this is true of any time period, but it's reasonable to state that it is especially true in today's markets. and i wholeheartedly agree that this introduces bias in the overall sample that should (and can!) be controlled for. (most notably Heckman won a nobel prize in economics for introducing empirical methods to deal with these issues.)

so i do agree with you that there is bias in the sample overall. but the question of the median and the distribution being skewed is actually a different point altogether. when a distribution is skewed, the common use of the word is that one of the tails of the distribution shoots out further (to the right, in the housing case) than the other. the median almost by definition is therefore less susceptible to skewness than the mean.

but really, maybe there's no point in getting technical. i still do stand by the fact that using a trumped up example of n=5 is quite silly, and that in a sample of 1000+, the median is still an informative measure of the distribution's "location" that is less susceptible to misinterpretation from skewness.

Juiceman and DTY, I take your points. The median ppsf is more reliable and stable than the median price. But luxury units, i.e. those more likely to be purchased with more cash, have meaningfully different ppsf than middle-market units. This cannot be ignored or downplayed. Sounds like I used the word "skewed" wrong, but I don't think the fact that the data sample is systematically biased is not a "technical" point". Market watchers like urbandigs and Jonathan Miller have made much of this point.

Sorry, having trouble posting.

Juiceman and DTY, I take your points. Sounds like I used the word "skew" wrong! Sorry. But I don't think it's a "technical point" that the median is a biased sample. I think it is meaningful in understanding the current market numbers, and urbandigs and Jonathan Miller have covered this point at length more eloquently than myself.

Juiceman, I agree that the median ppsf is a more stable, reliable number. But luxury units have meaningfully higher price per square feet than middle market units.

This is the thing: for close to a year, there have been screeching sound about the prices in NYC crushing. A year is a long time. Can it be shown NOW that those screechers had merit? Some prices adjusted down some up. Nothing remotedly resembling crushing. Those who have been making those sounds continue to rent and will be doing so forever.

JuiceMan, I'm in on the bet. I'm guessing over as well. dco, the charity idea is a great one - I don't see what you're worried about here.

Inquirer, I don't want to defend the shrillness of some of the naysayers. I suspect that even if the bias in the data were adjusted, there has not been a crash like many have been predicting. You are right on that point. As market-timers, many naysayers were very wrong about the fall of the NY real estate market. From a value investing perspective, though, I'm not sure they were wrong. They were and are arguing that historical relationships between prices and income, or prices and rents, seem to be out of whack in a way that has inflated RE values. Just as Warren Buffett doesn't buy a stock that is overvalued even if it might continue to rise in price, some potential buyers will not buy even in a rising market -- because they think RE is overpriced. (Please, I am NOT saying that the naysayers are mini-Warren Buffetts!) I believe that real estate is a market, like any other, that sometimes gets out of whack and will get back in line over the long run. So, I am just waiting for that to occur. I think there is a lot of data to suggest that this price correction will happen over the course of the next 3-4 years. I am not in the 50% or 40% camp, but I believe 30% peak-to-trough in real numbers is likely. It will happen gradually over this time frame, not in a quarter, and I don't expect it to start until Q1/Q2 2009.

TenthStreet-= Welcome and don't hold back. Even if you think the market will fall 5%, you will be ridiculed by the "NYC RE. Marketing Machine" also known as the "crew". I'll let you figure out who they are, on your own. Good luck.

"But luxury units, i.e. those more likely to be purchased with more cash, have meaningfully different ppsf than middle-market units. This cannot be ignored or downplayed."

I don't understand TenthStreet. The median doesn't ignore or downplay this. If we are comparing two periods of time a year apart, it would require a tremendous shift in the sale upscale (or sale of downscale) units to impact the median. A tremendous shift in upscale or downscale sales a year from now vs. today would result in the naysayers being wrong anyway.

However, if any of the bears would like to change their predictions to 30-40% correction of units 2 bed and below, or in new developments only, or with a current psft of $1200 or lower, I'm all ears. So far, I have heard nothing of the sort. How about this, today the median price per square ft for units under $1.5M is $978. 30% reduction would be $685 on 7/22/09. Would an upper bound help people stand behind their claims? Surely if we limit the sample to only the low to mid end apartments that would ease your median fears?

dco, no one is ridiculing TenthStreet, why don't you answer some of these questions?

Tenthstreet: "I thought Steve was wrong at first also. He's not, but it takes a minute to figure out why...."

You are correct.

JuiceMan: "we are talking about median price per square ft and not median price, so there is less of a bias."

Actually, to prove that you would have to take the universe of apartments in Manhattan and calculate how a particular quarter's data correlate to that universe, and normalize. You don't have the data to do that.

"the median almost by definition is therefore less susceptible to skewness than the mean."

Almost, and usually, but unless I could calculate the probability of that (not) happening, I wouldn't bet on it.

Of course JuiceMan and others don't believe the Fed because they prefer their own models because they prove their points, but that's why the Federal Reserve normalize their data:

www.ny.frb.org/research/staff_reports/sr218.pdf

and you will see that the data are normalized because units can't be compared across markets. Nor can units (in a small sample size) be compared if they are materially different (luxury penthouses vs. rent-stabilized Inwood, both included in the Manhattan streeteasy data).

Dudes - you're betting. Name me one successful gambler who doesn't know the odds of winning before he makes a bet. Name me on successful investor who doesn't know the downside risk (you quants out there, you!).

You can't. This bet is stupid.

Juiceman, in normal conditions the median shouldn't be too volatile. But in current market conditions, it is more likely. Per the Prudential Q2 '08 report, sales volume is down 21.8% YoY. That's a fair amount of volatility, and makes the median less reliable. Imagine the Q2 '07 numbers with 21.8% of its trades removed (with a disproportionate share of those removals being lower-ppsf trades), and the median will materially overstate the "true" median of all units in Manhattan. It's just a limitation of the data. That doesn't mean it has no value, or that it's dramatically wrong, it just tells me that the current data might be masking the softening of the market.

Yeesh, for all the baiting and taunting you do on this board stevejhx, you won't put up a measly $100 (measly in terms of your average stated income) for charity in the name of your predictions? I know, what's being proposed isn't adhering to the very strict definitions you desire, but it's entirely reasonable, and you seem quite timid about what amounts to a pretty friendly bet here. As I've told you before, I admire and even agree with a fair amount of your analysis, but your attitude has become frequently off-putting, and concluding "this bet is stupid" really doesn't do much for your credibility here. I'm not rooting for anyone here - I'm just a bit stunned that you've devoted so much time to posting here (and posted with such defiance and confidence) yet won't risk a relatively small amount of cash for charity.

"and you will see that the data are normalized because units can't be compared across markets. Nor can units (in a small sample size) be compared if they are materially different (luxury penthouses vs. rent-stabilized Inwood, both included in the Manhattan streeteasy data)."

You are right steve. You can't compare NYC to Miami. I've been saying this for months. You however, compare your Chelsea rental to every market in Manhattan. If what you posted above true, how can you do that? LMAO

Paging Mr. Backpedal. Paging Mr. Backpedal. George Bush is on the phone looking for advice on how to avoid claims, duck questions, and take responsibility for his actions. Paging Mr. Backpedal.

bjw, I didn't read the charity part of the bet. Somebody please capsulize.

"You can't compare NYC to Miami."

Because, JuiceMan, I have always stated that the reasons the markets were out of equilibrium in Miami and NYC were different, but the fact that they are out of equilibrium remains. Go back and read my posts.

Basically, JuiceMan, the mean price-to-rent (or imputed rent) RATIOS are different in each market. So let's say that in Miami it's 10 and in Manhattan it's 14. Just for fun's sake. You then need to normalize for that difference when comparing across markets.

Absolutely nothing to do with what I've been taking about. Check out the back tables in the Federal Reserves' "flawed" white paper and you will see what I mean. The graphs, too.

No backpedaling. You're just not bright enough to understand.

TenthStreet, I understand what you are saying but I disagree that the result would be material. You are assuming that the 21.8% sales volume is coming from the lowest psft units, but if you listen to the arguments on this board, new developments will get hit the hardest, which are the highest psft units. Either way, the argument that the median is not an accurate measure of real estate performance contradicts the "high end apartments skew the median" statement. If lots of high end psft places are selling, how in the hell can we have a 30-50% blanket reduction?

What I find interesting is that no one can come up with a statement for where they feel a 30-50% reduction will come from. If they feel it will be in the low end of the market, than let’s measure that. High end, then we can measure that. 1 beds? Chelsea? All we have right now are blanket 30-50% predictions. If you are going to use a blanket 30-50% prediction than we have to use MEDIAN PRICE PER SQUARE FT. It is that simple really. Forecasters need to either qualify their prediction or accept the measure. TenthStreet, do you understand what I’m saying?

stevejhx, one of JuiceMan's earlier posts in this thread:

JuiceMan
about 20 hours ago
ignore this person
report abuse Surely if you are confident enough to make 30-40% predictions you should be able to put something behind it?

ok dco, can we make it a little interesting? How about this, we use streeteasy median prices as of 7/21/08 and 7/21/09, "must have address box" checked, and a 30% reduction off median prices ($825 a sqft). The winner picks a charity and the loser donates $100 to it. Anyone on this board can play. Just respond to this thread with either under $825 psft or over $825 psft median target and the dollar amount you are pledging. Next year, those who responded correctly will pick the charity and those who chose incorrectly will donate their pledged amount to it.

Now dco, surely you can come up with $100 for a worthwhile cause and some bragging rights?

"You're just not bright enough to understand."

Tisk, tisk, tisk steve. Personal insults? Well steve, I'm bright enough to debunk your bull in my spare time and it's not even challenging. I would let you carry my bag, but you probably couldn't lift it.

bjw, just for fun you can count me in.

However, the bet is flawed because there is an enormous amount of new inventory coming online in the next year.

But I'll do it anyway - provided that it excludes "listings already in contract."

Here are the current data:

Sales in Manhattan
We found 7,480 listings with an address
Median price: $1,275,000 Median size: 1,176 ft² Median price per ft²: $1,169

So let's shoot for $820 psf.

Funny how the only ones unwilling to bet, and doing all sorts of contortions to come up with reasons why they won't, are steve and dco.

Juiceman, I understand what you are saying! The prediction I made stands: a real 30% peak-to-trough decline in median ppsf as published by Prudential Douglas Elliman starting in Q1/2 2009 and lasting for three years. I will not water this down or try to hedge it.

Also, here's my logic on the way things are going to unravel. I believe we are now seeing lower ppsf units not trade, as these are income / financing-dependent buyers. This temporarily skews the median (and the averages) upward (while volumes fall). Eventually luxury units will decline as well. Once everything is going down, the median will too.

I thought this chart would be a nice addition to the conversation. I'm not sure what I like most, the fact that it says "price per square foot has been relatively stable for the past three years" or the fact that we are using the MEDIAN price per square foot, which would be even more stable.

http://millersamuel.com/charts/gallery-view.php?ViewNode=1216608024OgigB&Record=6