Some additional numbers on the twelve properties the Lembis put on the market in May:

Lembi hopes to generate $43 million in revenue from the disposition, which equates to approximately $185,000 per unit, approximately 50% of replacement cost and a significant discount to what it paid to acquire the properties.

The buildings may sell as a group or individually. Most were constructed in the first quarter of the 21st century and a few were built in the 1960s. The per-unit prices range from $320,000 to $100,000. Most of the projected cap rates on the buildings are in the 5% range and are based on scheduled income and a 3% vacancy factor. Much of the interest so far has been local buyers each looking to purchase one or two properties, according to local sources.

The sales will be watched closely by the market because only one other comparable property has sold in San Francisco this year, according to Real Capital Analytics. The property was Empire, a 40-unit, four-story property built in 1907 at 1040 Leavenworth Street. The property sold for $5.8 million or $145,000 per unit; the pro forma cap rate was 4.7%. All of the Lembi properties are said to be of higher quality.

According to MPF Research, San Francisco rents dropped 5.2% in the first quarter of 2009.

65 thoughts on “A Few More Numbers For The Most Recent Lembi Twelve”
  1. The Lembis purchased $1.2 billion in SF apartment buildings between 2003-2007. From what I’ve read, many of these buildings were purchased with very little equity (as high as 92% LTV), at very low cap rates, making the investments largely cash flow negative. Here’s what I want to know: Why would anyone care to bid on the first 12 buildings, in what will is likely to become a San Francisco multi-family property puke-fest? 5% cap rate? No thanks. Call me when it get to 10%.

  2. 5.2% rent drop in the first quarter sounds about right. From what I’m hearing from a few recent transplants I know that moved into the city, both for grad school, rents have fallen a lot more in the last couple of months. With unemployment rising and population leaving SF, I bet rents continue to drop faster and farther.

  3. ^ my rentals went up over 20% in the last 2 years. As of now I’m down 5-8%. Still ahead of the curve, and no, I don’t expect rents to go below 06-07, at least in my case. Don’t hold your breath for wholesale and massive rent reductions (sans specific situations.)

  4. Do not mix up a cap rate with the overall return for a property. With most small investment properties in the Bay Area over the last several decades, the majority of the return has been made up by asset price inflation, not by the current income which the cap rate measures. And waiting for cap rates to double is like waiting for house prices to halve – pretty unlikely.

  5. I wonder how much deferred maintenance you get with that 5.0% cap? From the Citiapartment buildings in my neighborhood, most look pretty poorly maintained.

  6. ^ dude, that mash up is based solely on asking rents. My numbers come from actual rents in the last 3 years. But details…

  7. Wow, Miles, breaking out the “asset price inflation” mantra?!?! How quaint! Um, 2005 called and wants its comment back. 🙂

  8. “My numbers come from actual rents in the last 3 years.”
    Well, your numbers are just one data point. Craigslist, on the other hand, is *the* market.
    I don’t know about reduction on your current address, but you can expect a wholesale and massive reduction if you are willing to move. Archstone Fox Plaza, for instance, advertised $1800 for 820 sf that is listed as $2800 on their site just last weekend.

  9. Well, your numbers are just one data point. Craigslist, on the other hand, is *the* market.
    I would point out that while Craigslist transmits a lot of information, you have to keep in mind that
    1) It contains only asking rents
    2) If the asking rent is too high, it sits on craigslist and is constantly republished, possibly with lower asking prices each time.
    3) If the asking rent is too low, it quickly disappears from craigslist.
    Because of this, by design, Craigslist overstates rents.
    By how much? Well, we have data for rents paid via the census. I.e. in 2007, we can look at the median rent paid by new tenants (e.g. moved in within 1 year), grouped by # of bedrooms, and compare this to sf rent stats.
    Since the census data includes people who moved in within a year of 2007, and there seemed to be an asking rent price “bubble” around mid-year of 2007, which popped in the beginning of 2009, let’s be genereous and only compare the ACS data with the data form October 06 to May 07. In this case, we have (actual) paid rent vs craigslist asking rent as follows:
    bedrooms | ACS median | Craigslist | Difference
    ——————————————
    1 | 961 | 1770 | +84%
    2 | 1518 | 2500 | +65%
    3 | 1923 | 3000 | +56%
    So, you see, there is a huge difference between the craigslist median, and the actual rent paid — even for new tenants — so no rent control issues.
    One thing that would really help here, would be if craigslist tracked how many times a unit had been posted, with the price of each listing. This way, you could track it’s price decline until sold, and then use only the final (sales) price to gauge the market.

  10. Btw, 45y.o hipster mentioned 3 year median prices, so I thought I would reproduce those here, with hhincomes as an added bonus. Of course, these are ’05-’07 — but relevant for the time period in which the bulk of these properties were purchased.
    census=# SELECT bedrooms, median(rent) AS rent, median(hhincome) AS hhincome FROM pums WHERE (city = 6290 and year >= 2005 and movedin = 1 and rent > 100 and hhincome > 100) GROUP BY bedrooms ORDER BY bedrooms ASC;
    bedrooms | rent | hhincome
    ———-+——–+———-
    1 | 947 | 26463
    2 | 1409 | 60372
    3 | 1821 | 77914.5
    4 | 2119 | 82104.5

  11. Robert, is there some program you use to extract/load the census data into an SQL database, or just something home brewed yourself? I’m a fellow data junkie, and the census data seems very interesting to mine. Thanks!

  12. With the over-hanging of this new proposal to cap rent at 1/3 of income, why would anyone want to become owners of large apt buildings at any price??

  13. Robert, Wouldn’t that data include subsidized prices on new places? BMR units and other rental places operated by non-profits and the MOH aren’t somehow broken out from census data, are they?

  14. ^ I was also wonderin’ about that, cause $1518 for a 2br and $1923 for a 3br is much lower than the actual citywide numbers. I’d venture to say that most market rate 2br’s range from $2000-3000 (from the outer sunset to PAC heights.). But the point still stands that the craigslist data of asking rents, especially for evaluating the delta in recent rental declines, is very inaccurate.

  15. “So, you see, there is a huge difference between the craigslist median, and the actual rent paid”
    True. In fact, I just signed lease with 15% less than the posted price. Not to mention things like 2 mo free offerings not captured in list. However, as long as the comparison is apple-to-apple, the craigslist should provide some information about the trend.
    There is more room for negotiation now than 2007-8, no doubt. Which means, the drop in actual rent should be larger than what craigslist suggests. So, if anything, craigslist tends to understate the changes from peak to valley. Bottom line: there has been a massive rent reduction since the 2008 peak.

  16. ^ No! You can’t (accurately) state that across the board. I just gave you concrete proof to the contrary. It depends on many factors- some apartments fell more than others. But you just want to continue generalizing to fit your theory of massive rent reductions, massive price reductions, etc. If it makes you feel better…

  17. With the over-hanging of this new proposal to cap rent at 1/3 of income, why would anyone want to become owners of large apt buildings at any price??
    Renting at more than 1/3 income is riskier both for the renter and the landlord. Everyone should make his own choices in a perfect world, but the demand is such that way too many people will go over their 1/3 limit and go to 40%, 50% or even more. That in itself adds pressure to pricing and pushes many to follow the trend of rent overburden.
    On the other side, the 1/3 limit will have a gentrifying effect, with landlords cherry-picking renters even more. But if there are not enough deep pockets around, then that will cap rents somehow.
    There will be many unintended consequences to this decision I think but that is what happens when limits are applied.

  18. “But you just want to continue generalizing to fit your theory of massive rent reductions”
    Well, dunno about my “theory”, but yes, I am speaking generally for the market, and not your apartment in particular, ‘cos it’s true that rent has fallen massively in general since 2008 peak. And I think it’ll fall further for the rest of the year given the vacancy and employment situation.

  19. Gumby,
    You can download microdata from the ipums project:
    http://usa.ipums.org/usa/
    Fill out a request for your data set, and they will email you a link to download it, along with codebooks. I wrote a script to slurp it into a local postgres database. You have to be careful and read the handbooks and especially the field codes. For example, there is no “NULL”, so when something doesn’t apply, they will use a reserved value, such as 999999. If you don’t handle this, it will screw up your statistics. Have fun! Btw, I use PL/R for server side statistics, and python or R for making graphs.
    Brutus,
    No, subsidized housing doesn’t affect medians. For more info about medians and robust statistics, look in wikipedia.
    45yoh,
    Perhaps the editor deleted the posts where you proved something. In general, people over-estimate incomes in SF as well as rents. These are the median rents that people are actually paying, and so you have to do some soul searching on your own to make them fit with your perceptions. I can’t help you with that, other than to point that SF has a lot of hoods: Richmond, Sunset, Silver Terrace, etc, that don’t get a lot of mindshare, but that do affect medians. Also, most landlords here tend to be small time and not particularly savvy — witness the subject of this thread, so they tend to be deluded in terms of what the market can bear, and keep reposting asking prices that won’t find takers, which of skews all of these craigslist statistics.
    Fishtarian,
    Just be careful with craigslist. Even if you throw away the 90% + listings which are just re-postings of properties that don’t sell at those prices, you are still left with a measure of the margins. In other words, when craigslist asking prices shot up by 20% in the first half of ’07, rents paid did not shoot up by that amount. In the same way, when they collapsed by 20% in ’09, rents paid did not drop by 20%. Rents paid generally swim well below the craigslist rate, at all times.
    One way to understand this, is to think of the margins as the derivative of the rents paid — so it will be more wild. The actual rents paid is the sum total of negotiated rents for a given time period. If, as is the case, 25% of renter households move every year, then you can be sure that 1) they are more likely to move in when asking rents are lower, and move out when asking rents are higher, adjusted for incomes. Moreover, 2) It would take 4 years for 90% of the pool to have moved. Finally, the median household continues to pay about 25% of their income on rents, so if you want to calculate median rents, then income data is a fast and easy approximation to do it. So, really, there should be no argument as to what is being paid, since income data is easily available. Again, it takes some soul searching for some on this board to accept the income data.

  20. ^ and so you think that this cenus data is accurate? No way the average rent for a 2br in SF in 2007 was $1518. Does that incl. Section 8? What about other subsidized rents? This definitely does not represent market rate rents, which is what most of us are interested in.
    And what I said earlier ‘proves’ that not all rentals took large reductions. An apples to apples for rents, if you will. You’d think people prone to making wild exagerations (of the massive rent reduction variety) would stop and think about anecdotal data as well. But apparently not if it doesn’t fit their bias.

  21. 45 y.o.
    First, understand what a median is, then re-read my post and respond to the arguments I made. The only non-sensical statement in the above is the questioning of the accuracy of census data, which of course is the most accurate data we have.

  22. “One way to understand this, is to think of the margins as the derivative of the rents paid”
    What data to use depends on your purpose, I suppose. For my purpose, the “market” is not all rental units, which census data probably capture more accurately. Rather, it is the units currently on the market. And it’s probably true that craigslist may not be accurate since there are bigger discrepancies for negotiated rent in softer markets, but craigslist data is what the vacant units are asking for at any given time.

  23. census data, which of course is the most accurate data we have
    If those numbers look low to 45YOH as an owner, they look low to me as a renter.
    Is the percentage of subsidized housing known for that census? I know a few wouldn’t change the median, but is there a chance that subsidized units are somehow over-represented in the data? Or perhaps much more common than expected?

  24. “In the same way, when they collapsed by 20% in ’09, rents paid did not drop by 20%.”
    This is true. But if you signed the lease at the ’08 peak and looking to renew or looking for a new place now, it should be reasonable to expect to pay 20% less.

  25. Robert,
    Are we not attempting to determine median actual rents on market-rate properties? I thought that’s what we were looking at and interested in. If we’re simply looking for median of all properties, then yes, subsidized housing wouldn’t matter. Since the average person looking for a market-rate property in SF does not look at subsidized housing, that would seem to be an entirely different market, but doesn’t seem to be broken out in census data.
    If I’m missing something obvious, let me know, instead of just being a jerk and directing me to Wikipedia.

  26. The census data will include all the people who have rent controlled apartments they rented 20 years ago and are now paying way under market rents for.
    Census data doesn’t tell us anything about what the market rate is for rents today. At least not in San Francisco.

  27. NVJ,
    Did you skip over the part where I only included those who have moved into their apartment in one year? That pretty much destroys your argument, right?

  28. ^^^Robert is only looking at data from people who have lived in a place a year or less, which would remove rent control distortions (outside of a random few people who gamed the system – became a roommate on an old lease, etc).
    However, with almost all new construction of the past few years falling into two categories:
    1. Market rate for-sale condos
    2. Below market rate and subsidized rental apartments.
    I’m not convinced that we’re seeing a good picture of market rents. The TNDC alone had 250 new (never before rented) rental units hit in 2007 which would all fall at the very bottom of the range. I’d have to know turnover rates for the different types of housing and the total number to know if the number of subsidized units is enough to make a large difference to the median, but I suspect that it’s at least 15-20% of the total rental units moved into in 2007.

  29. Brutus,
    Unless you are claiming that more than half of renters in our universe are paying subsidized rates (recall, the universe is those paying cash rent who have moved in within one year), then no, it wont have any bearing on the median. The thing about medians is that they are robust even when up to 50% of the data is bad.
    Of course, if you are claiming that more than 50% of renters who moved in within one year and who are paying cash rent are receiving subsidies, then this would be an interesting point to make, and would require evidence. Merely asserting the existence of subsidies is vacuous when discussing medians.

  30. Robert, if the subsidies are bringing in people who would not be in the market that we’re attempting to look at without the subsidies, then it is affecting the median that we want to look at it. If we’re assuming that all of these people and units with subsidies would simply be charging a little higher without them, then it’s a different story. 50% of the data can be bad, but if all of the bad data is located at one end, it will cause massive distoritions and make the median less robust, no?
    If you’re trying to find the median price of oranges, you can’t look at a data set that includes apples and oranges and expect it to show you something.

  31. To add to the discussion regarding subsidies, keep in mind that if 1/4 of renters move every year, and you have 200,000 renter households, then you are talking about 50,000 households who have been in their apartment for a year or less.
    So, you would need more than 25,000 households receiving subsidies, and *all* of these would need to be *new* arrivals.
    Of course, if a household is receiving a subsidy, they would be less likely to move. Certainly, there is no reason to believe that the distribution of subsidies would be skewed to new arrivals. So, you are really claiming that half of all rental households in SF are receiving non-cash rental subsidies. If so, I really would like to see evidence of this.
    Moreover, such a view is inconsistent with the data, which has been fairly constant over time. Specifically, for new arrivals, take their median income, and multiple by about 27%, and you get the median rent paid.
    This has been the case for every census since 1980, and every ACS survey ever taken for San Francisco. That figure is remarkably robust.

  32. Robert, if the subsidies are bringing in people who would not be in the market that we’re attempting to look at without the subsidies
    Well, you would need to make the point that throwing away the bottom half does not throw away the group that is receiving the subsidy, and in fact that the subsidy receiving group materially affects the median.
    Go ahead, and make this case — using data — it’s a bit of a tough case to make.
    And be sure to explain why the median rent is consistent with rents paid in the rest of the country, as well as the rest of the bay area, as well as historical data for the country and the bay area. I.e. abut 1/4 of the median household income is the median rent paid. You would need to argue how a portion of the population receiving subsidies has materially affected this, when no change in rental burdens has been observed.
    In all seriousness, I would be very interested in seeing data that could pass through all of these hoops.

  33. 1/4 of renters move every year? That number seems absurdly high for SF, but if that is what it is, you’re probably correct.
    Of course, if a household is receiving a subsidy, they would be less likely to move. Certainly, there is no reason to believe that the distribution of subsidies would be skewed to new arrivals. So, you are really claiming that half of all rental households in SF are receiving non-cash rental subsidies. If so, I really would like to see evidence of this.
    The reason that I would believe that it is skewed to new arrivals (not necessarily new arrivals to SF, but new arrivals to housing of their own in SF) is that so much housing has been added to the supply over the past few years to ONLY be used for these folks. We haven’t added subsidies to existing housing. We’ve added housing that wasn’t a part of the supply and can only be rented in a subsidized manner. Each year that supply is only added at the bottom, the median becomes more and more detached from the market rate figures.
    In a bifurcated market like SF, I would think that would matter a great deal to the discussion of market rate rents.

  34. Well, you would need to make the point that throwing away the bottom half does not throw away the group that is receiving the subsidy, and in fact that the subsidy receiving group materially affects the median.
    I would say that it materially affects the median because it’s housing that would not have been built without a subsidy. It would not exist without the subsidy. The people living in the housing would not live in SF without the housing being built (or at least would not live in their own housing). It’s not a case of these people paying slightly more, which would still be in the lower half – it’s a case of adding units to the bottom that are not REALLY a part of the SF housing stock, as far as determining rents are concerned.
    Now, again, if there really are 50,000 units that turnover each year, the subsidized housing numbers probably don’t matter much. There probably haven’t been more than 3,000 new units of the housing that I’m talking about built over the past five years.

  35. 1. why are you so sure that this census data is so accurate? it’s not as if the gov census data is not prone to errors, omissions, bias, etc. you assert it’s very accurate. i’d like to see more info supporting this.
    2. as you saw from others, the interest is in market rate rents. and it does not take a PhD in statistics to know that a median value is affected if a significant subsets of it’s data is erroneous (i.e. subsidized housing or corporate rentals.)
    lastly, your condenscending tone is unnecessary, for it belies your inability to look at reality (facts on the ground), for there is no friggin’ way that the average market rate rent paid in SF in 2007 for a 2br is $1518. marcus and millichap run average rent stats on a regular basis. and being as resourceful as you seem to be, i’m sure you can find other organizations that track SF rents. it doesn’t take a PhD in stats to realize that the numbers you presented (for market rate) are inaccurate. sorry bro, but this is not passing the SF landlords smell test.

  36. The reason that I would believe that it is skewed to new arrivals is that so much housing has been added to the supply over the past few years to ONLY be used for these folks.
    Now, we are having a real discussion!
    Re: mobility of renters:
    I have to say, I made an error with the decay — it’s exp^(-.25) not .25, so after 1 year, that would be about .78, or 22%. That, to be fair, is a regression from aggregate ACS data (not microdata), but the R^2 is 0.997.
    Alternately, I just ran a query on the microdata, and the result is also 22%.
    Finally, I found a new ACS table — B07013, which says there were about 402K renters in 2007, and 317K of them lived in the same residence 1 year ago, so that would yield a figure of 21%.
    Note that renters, in general are far more mobile than owners. And san francisco is a more mobile city than the nation as a whole.
    As an aside, this is why rent control has failed to affect median rents — there simply aren’t enough old timers to affect medians. Not to say that many landlords are stuck with below market tenants, but there are too few of them to change the median from the historical pattern.
    In terms of new housing stock added, the short answer is that not much is being added. In table B25034, it says that about 3K *units* were added from 2005-2007. So, that would be about .08% of the total housing stock in a *two* year period.
    Of course, this includes all housing stock — for both owners and renters, both above and below market rates. So, I don’t see how the numbers add up for there to be enough new below market units to affect medians. Note only 7% of the housing stock was built before 1990 (2007 data).
    Here is a link to these and other tables.
    Now about the subsidy issue. If we assume that 1/4 of the housing built since 1990 was below market rental housing, that gives us about 1.75% of units, or about 3500 units. The SFHA is “authorized” to provide for 12,000 public housing units and 21,000 section 8 participants (not households). Although, I’ve read elsewhere the number *funded* is 6,000 households. In any case, let’s overcount:
    12,000 *households* for section 8
    3,500 BMR households
    21,000 “project” units.
    for a total of 36,500 households with subsidized rents — or about 18% of the total. This of course assumes there is no overlap.
    In any case, one approach would be to just ignore this population before taking the median. But these programs are means tested, so lets trim the bottom 18% of household incomes and run the computation again.
    bedrooms | median
    ———-+——–
    1 | 1019
    2 | 1509
    3 | 1923
    Notice that the median shifts only a little, i.e. between 7% and 6%, when we drop the bottom 18% of the sample.
    Another analysis would be to drop only the public housing, or 10.5%, corresponding to those earning about 18K or less.
    The results of this query are:
    bedrooms | median
    ———-+——–
    1 | 1012
    2 | 1426
    3 | 1834
    Notice that these numbers are basically the same as the full trim. So, if it makes a difference to you, you can take any of the three samples provided — all of them show massive differences between the craigslist *asking* price and the rent paid.
    However, I would point out that there are serious problems with trimming *only* the bottom of a distribution. Specifically, you are ignoring:
    1) Setting aside scarce resources (housing) for some does not drive up the costs paid by others — leaving the median unchanged, or possibly skewed higher. In particular, you are ignoring that the cost of the BMR units is borne by other renters, and is not “lost” revenue.
    2) You are invalidating the low outliers but not the high outliers, such as temporary rentals, corporate rentals, insurance rentals, etc. For this reason, usually when you trim, you trim both the top *and* bottom of a distribution, because the “median” household is not in the projects, nor are they renting a temporary unit, etc.
    In any case, it is a fascinating discussion, and I think the point about the massive over-statement of rents by looking at craigslist still stands, and that the reason for this overstatement are the dynamics I outlined previously in this thread.

  37. I think you are ignoring the considerable amount of subleasing that goes on in this town. Just go and look at the number of “rooms for rent” it is almost half of the total number of units for rent.

  38. Well, 45yo hipster, I don’t know about average rents — you seem to confuse average with median — but in my experience “on the ground” — these numbers are about what people I know paid — those who moved in during that specific time period of mid-06 to mid-07.
    My next door neighbor got her place for about 1550 in ’06, but Noe is more expensive. I would add that I pay less than the ’06 medians now, but I moved in earlier than this. A friend of mine just got a 2 bedroom at 18th and Castro for about 1500 — she’s moving in in the beginning of July. Note that except for my neighbor, the rent paid was less than the advertised rates. In my case the landlord offered me less, and in her case, she negotiated $200 less. These are all 2/1s.
    I know these are anecdotes, but I brought it up since you questioned my “on the ground” knowledge. I am, after all, a renter, and most friends of mine are renters. I know what you can and cannot get, after all, provided that you are a little flexible and don’t require a deck, or a yard for a dog, an in unit washer, etc.
    Of course, these are just examples, and I make no claims about medians from them — I provided the median data elsewhere.
    The condescending tone, is of course shared, but to be honest, for my part it stems from a certain lack of reading comprehension and logic that I detect in your posts:
    You vociferously kept pointing out that there exists some units which charged more than they did in the past, and this somehow “proves” that rents are not falling. Well, to be honest, that’s a pretty vacuous statement, don’t you think? I mean, no one is claiming that *all* units are charging less. And to keep pounding on this strawman, in the words of Tito, “Like a drunk hanging onto a fence” is not adding anything to the conversation.
    This is of course more ironic since I was not arguing that rents were falling at all, merely that craigslist overstates the rents paid. If I were arguing that rents are falling, I would have provided data for median rents in 2000, and compared that to data in ’07. But I compared ’07 data to craigslist data for the comparable time period.
    Now, I may believe that rents paid will fall in the future, but what we have now is a fall in asking rents. And I specifically said that just as the rise in asking rents was not matched by a corresponding rise in rents paid, so the fall is not so match.
    Moreover, I provided hard data for this, as well as a rationale as to the underlying dynamics. But, you were unable to respond to either my arguments, or my conclusions, and instead attacked only a strawman throughout the entire thread. Hence the gripes and the tone.
    What would be nice, is if you actually parsed the arguments and provided some thoughtful analysis:
    Why would the census be lying? What would distort the picture? Why would craigslist *not* overstate rents — seeing as units that fail to sell are reposted, whereas units that do sell are withdrawn? You can really get into an interesting discussion with a little bit of logic and analysis. I am not always right, and am ready to admit when I am wrong, but I am able to understand what someone is saying, and respond to their actual arguments.

  39. Well, NVJ, this is a good point — subleasing is real and affects medians.
    But, this highlights the case rather than undermines it. I.e. there is a choice between subleasing and renting, so the rents paid should measure subleasing; we are not trying to measure how much money a landlord would get. Maybe the landlords on this thread were thrown off by this. We are measuring how much money a “new” tenant needs to pay to live in this city. I.e. rents paid — it is a different measure.
    However, for the census data, an interesting thing happens when you look at household records. When they combine the household records from person records — because a household is not a family — if different members of the household have different tenure, the “movedin” field is set to “n/a”.
    I only queried the movedin field to be 1, or “within the last twelve months”. Because of this, my universe really consists of entire households that have lived in the unit for 12 months or less. BTW, this can (and often does) also involve subleasing. I know of several “illegal” subleasers, and in fact it is quite common.
    Now, if you throw in the “n/a” field for moved in — then you cannot look at recent renters, and rent control comes into the picture — of course the numbers are smaller.
    But you are right, certainly subleasing is one more reason why there is a big difference between asking rents, and rents paid.

  40. Mea Culpa, NVJ, I was wrong about combining the records.
    movedinmeasures when the “householder” moved in.
    “The information for 1960 and 1970 was taken from individual person records, while that for 1980-2000, the ACS, and the PRCS was drawn from the household records. The year that the householder moved in is not necessarily the same year other members of the household moved in. Non-householders in 1980-2000, the ACS, and the PRCS are coded N/A. ”
    The “householder” (formerly known as “head of household” is defined as:
    “The householder refers to the person (or one of the people) in whose name the housing unit is owned or rented (maintained) or, if there is no such person, any adult member, excluding roomers, boarders, or paid employees. If the house is owned or rented jointly by a married couple, the householder may be either the husband or the wife. The person designated as the householder is the “reference person” to whom the relationship of all other household members, if any, is recorded.”(emphasis added)
    It does seem to me, though, that a movedin value of 1 should require either that the unit be rented within one year, or possibly an illegal subleaser, in which the original tenants are gone, but the current householder was put on the lease. You can read the census definitions and draw your own conclusions.
    Time for bed.

  41. “I think the point about the massive over-statement of rents by looking at craigslist still stands”
    Robert, your point of rents being more stable than what’s advertised on craigslist is well taken. If we separate “rents” from “market rate”, then I think we are in agreement. Craigslist *is* the market and therefore represents the market rate, albeit the actual negotiated market rate may be less. It is a useful tool, as long as we are comparing asking rent to asking rent as SFRentStats does. In a real time, no less.
    Anyway, This has been a useful discussion for me and I’m glad I kicked it off by barking at 45yo’s assertion that there has been no massive reduction in rent. And I should correct myself here that there has been a massive reduction in the market rate of rent, rather than rent.
    Gotta get back to packing, I’m moving from ORH to Paramount today…

  42. Interesting discussion about rents and stats.
    My experience with renting is that the best bargains seem to be in the SFH market. ~25% off craigslist wishing prices is what I’ve been able to get both times I looked (2002 and 2008). Never had even a dime of rent increase, and we pay less today than in 2002 (but not totally an apples to apples comparison, as they are different places), and in each case less than we would have paid in 1997-2000 if we had been here then.
    I’m not sure why people continue to “rent” these sorts of houses from the bank at all-in costs of $6-10K per month (not counting the depreciation of purchase price which is a certainty at these levels), when they are available to rent from long time owners for $3-4K.
    I’ve heard all the arguments (stability, ability to “change the colors”, etc.) and if that is worth the additional $3K+ per month (plus the selling cost and capital loss when you sell) for people, that’s fine. I suspect, however, that most think they are making a “wise investment”, when in reality they are just paying 2-3x for a consumption good with significant illiquidity penalties, a fact that is becoming more obvious as the bubble unwind unfolds.

  43. I’ve heard all the arguments (stability, ability to “change the colors”, etc.) and if that is worth the additional $3K+ per month (plus the selling cost and capital loss when you sell) for people, that’s fine. I suspect, however, that most think they are making a “wise investment”, when in reality they are just paying 2-3x for a consumption good with significant illiquidity penalties, a fact that is becoming more obvious as the bubble unwind unfolds.
    You have heard all the arguments, ad nauseum. Many of us have also heard all of yours. I do not see the point in recycling these arguments constantly.

  44. Regarding “changing the colors” I recall painting two or three of my rented apartments while in college (often because the existing paint was old and dingy) and never heard a peep of complaint from the landlords. We usually forfeited about 25% of the damage deposit for various other reasons, but who really cares? — 1/4 of a month’s rent out of 2 or 3 years living in a place isn’t a big deal (that raises your costs by less than 1% overall).
    I also planted a vegetable garden at one place, hung up curtains, pictures, shelving and all kinds of other wall-damaging items pretty much whenever I wanted to, installed a dishwasher once (myself), but as long as the rent was paid on the 1st every month I never heard a single word of complaint from the landlords and never once asked permission.

  45. anonn, you have seriously become nothing but an old, bitter fart contributing nothing but mockery of those who actually take the time to provide thoughtful comments. Jimmy (NLB)’s follow-up to LMRiM’s post confirmed that the extreme over-valuation of the benefits of ownership is a valid and non-obvious issue to note. It puts the lie to a common RE industry sales pitch.
    I really think you need to consider a new profession. This downturn has obviously caused a substantial hit to your earnings and it appears that you’re losing it.

  46. ^^Uh oh. That can only mean one thing:
    As soon as you see it, run to SF Schtuff!
    Condo and lofts: Down, down, down.
    Anonn’s favorite district 5: Volume off about 1/3, prices down by 10% from last year. District 7: Prices off by 20%, volume about half! District 9, incl SoMa: Volume off 40%, prices off over 20%!
    Things went from bad to worse last year. Now things are going from worse to catastrophic!
    http://sanfranciscoschtuff.com/2009/06/10/condo-and-loft-sales-may-%E2%80%9804-%E2%80%9806-%E2%80%9808-%E2%80%9809/#more-4600

  47. I happened to notice that the price on 2001 BMW Z8s has dropped from the $110k’s into the $80k range over the last year. A ‘green shoot’ in this recession? Quite possibly … I wonder if we’ll see $60k before long.

  48. oh robert, i deliberately responded to your posts en brut, or enfant terrible because deriving median rents from the census data is misleading FROM MY PERSPECTIVE (and i think from the perspectives of many others perusing this blog.) i see a limited value from the discussion above (sans an intellectual pursuit contained within a misleading dataset.) i agree with you that recent craigslist ‘asking’ rents are not correlating with actuals, and that has been discussed and corroborated here before.
    so, what question do i think is the more interesting and relevant one? an attempt to quantify by how much rents have been reduced in the last 6 months or so. perhaps we need to look at a tighter dataset so averages can be used to provide more meaningful information. if we normalize a definition for a ‘typical’ 2br rental in SF (i.e. 750-1000 sq ft, 1 ba, reasonable finishes, not sub terrain, no 7′ ceilings, no gov subsidies, no corp or short term rentals) than i’m quite confident that the $1518 median is not an accurate reflection of current market rents. this is especially the case for central SF (excluding D3 & 10; central/outer sunset & richmond), where we have the most rental units anyways.
    i am approaching this from a bottom up, apple-to-apple perspective with the info i gave (namely, that my rentals shot up 20-25% in the last 2 years, and are down 5-10% currently.) SOMA probably has larger reductions (where i think fishtarian derives his bias) due to the higher supply & recent sales-to-rental converts.)
    if anything, this discussion has illustrated that assembling meaningful data on rents is just as tricky and challenging as analyzing home values. as i mentioned in a prior post, there are other surveys that measure SF rent values (marcus and millichap’s comes to mind.) perhaps these could be brought into the fold, if you’re willing to venture out of the census data sandbox…and i promise to play nicely!

  49. Oh no, not the dreaded “it’s only down 5-10%” argument!
    Where have we heard that before? And how well did that work out for the people arguing that point?
    We just finished our college recruiting season. A typical year has our starting salaries at $X: this year they were at $0.8X. We give people 1 week to decide and they typically answer in about 5 days: this year, no one took more than 24 hours to accept. We usually get about 50-80% acceptance of our offers: this year was 100%. Last year, 100% of our new hires moved to their own apartments in SF: this year 100% of our new hires are living at home, outside of SF. Tell me how well that bodes for rents. And if my competitors offer their new hires more, I’ll undercut them on the price they charge their clients and I’ll steal all of their business.
    Rents look to be down about 15% already, and those are asking rents, which are negotiable for the first time in a long time. And come July, you’ll start to see a serious decrease in rents, as the normal outflow is supplemented by job losses, and there is very little inflow compared to years past.
    Ultimately, it’s people like me and other business owners who hand out the money to pay those sky high rents and this year, I refused to do it. I hired everyone I needed without it, thanks very much. You guys can just watch your apartments collect dust as they become vacant, there is a complete willingness by new college hires to live at home and they consider themselves lucky to have a job at all.
    One more stat: of my outgoing headcount, 100% are leaving the bay area. I wouldn’t invest in a bay area apartment building right now.

  50. Tipster — what kind of business do you run? Is it engineering? I’d like to cut my costs too; perhaps I can pay all my engineers and consultants 0.8X their current rates. That would be sweet.

  51. We just finished our college recruiting season. A typical year has our starting salaries at $X: this year they were at $0.8X.
    Hence the deflationary mindset. Of course, employers paying less in wages only means that their customers have less disposable income, so that sales decrease, so that wages are cut more, etc. Of course, first movers can benefit, etc, but the aggregate effect is the same.
    This is one reason why, in the great depression, the government tried to coordinate price increases by allowing industry cartels if those cartels agreed to unionize their workforce and pay higher wages. I don’t want to get into a fighting about the success of this coordination effort — I bring it up to point out that we can expect further coordination efforts if earnings continue to fall.
    Traditionally, the only way out here is an infusion of money. This could be in the form of Gold discoveries from Alaska, a la the 1890s, or a Gold Influx from Europe in the 1930s, or Emperor Augustus bringing crates of Gold to calm the financial panic of A.D. 33.
    In more modern economies, it will be in the form of deficit spending, or money from net exports. I’m not aware of any example of deflations ending “naturally” without an infusion of funds. In either case, employers need money that doesn’t come at the expense of cutting wages (hence reducing their sales), and consumers need money that doesn’t come at the expense of thrift (hence reducing their salaries).

  52. Interesting bit, Tipster. And a very micro view of a more macro sideshow: global wage arbitrage.
    I am not saying your newbies are being paid less directly due to globalization. But indirectly, through a complex chain of financial readjustments.
    For many years, we thought that the export of jobs would touch only the lower class. That all changed in the late 90s. For about a decade now highly skilled jobs are being exported/outsourced. We still retain a lot of good jobs, but some of the natural growth is being absorbed by others.
    This has caused an overall stagnation of salaries at home that was compensated with reckless borrowing. Now that we cannot (yet) borrow ourselves back into growth, the net result is wage deflation. And if we ever have inflation after the huge bailouts that will make wage-earners a bit poorer.
    What was bound to happen one day is hitting us now. You can’t try and compete with countries with low salaries and hope we’ll still have 4% wage increases forever.

  53. One data point:
    I live in a small-midsize building in lower Pac Hts.
    I occasionally watch Craigslist for ads for open units in this building.
    In the past (past 5 years at least) vacancies were filled immediately.
    In Spring 2008, a 2BR unit was advertised on Craigslist for $2500.
    This Spring, a 2BR unit was advertised on Craigslist for $2000. And it took over a month to move.
    The units are all the same. No price difference per unit.
    It’s one data point, but it’s definitely an apple: 20% drop in asking.

  54. ^ was $2500 the max for these units in the last 5 years? also, what do you think they were renting for spring of 2006?
    my experience was a 25% jump spring 06-08, and i dropped 5% and 10% spring 09 (based on actual signed leases.)
    my point, from an LL’s POV is that 25% jump came quickly, and it may go away (or partially go away) quickly. but overall rents are still quite high in SF. and people taking a salary haircut are also opting to share flats/rooms/etc. not all go back to live with mom and dad in da’creek (walnut creek.) the reality is always more nuanced, layered and complicated, thus rent decreases are contingent on many factors, that even a quant could not model with any degree of certainty. there are many forces at work.

  55. Ah, the old “rents won’t fall, everyone will just double up” argument.
    Kinda leaves out the answer to the question as to how far rents for the unoccupied other half of the apartments needs to go.
    An example: there is a city with two renters and two apartments owned by different owners. During a downturn, the two renters agree to share one of the apartments. They can go to each of the owners and get the best price because each owner knows the loser takes nothing.
    Prices fall dramatically in that scenario. The owners who tout this doubling up always seem to forget that minor detail! Demand is half, but the supply stays the same. Prices always fall.

  56. ^True tipster, but since everyone else is tossing in anecdotes:
    One of the units in my building was just rented by a couple who recently lost their house to foreclosure in San Ramon. They found that they simply couldn’t afford bridge tolls/gas/parking/BART/higher utilities and a place of their own in the areas of the East Bay that they liked. Their place here has higher rent than they would have had in SR, but lower overall costs, since both now have a ten minute Muni or bike commute, or a 30 minute walk.
    Of course, I’m not saying that means much. Just tossing out anecdotes 😉

  57. 45yo:
    “was $2500 the max for these units in the last 5 years? also, what do you think they were renting for spring of 2006?”
    I haven’t been watching that closely. I’m pretty sure they were getting $2700-2800 in 2006, but I could be wrong.

Leave a Reply

Your email address will not be published. Required fields are marked *