As we wrote in January:

The Marcus & Millichap rental outlook for 2009 in San Francisco: rents up 3.3% on 400 new units in professionally managed apartment buildings with at least 20 units.

Our comment (at the risking of stealing a bit of our own outlook thunder): we believe Marcus & Millichap is significantly underestimating both the number and impact of “shadow” market units for rent in San Francisco (which they deem to be “barely a factor”) as well as the effective number of new units for rent that will hit the market in 2009.

A new report from Marcus & Millichap, however, now calls for a 8.9% drop in residential rents in San Francisco by the end of the year, a dramatic 12.2 point swing in their forecast over the past six months. Our outlook and original rebuttal haven’t changed.
UPDATE: A plugged-in reader adds:

There have been more drastic decreases in SOMA. I wanted to move there last year but asking rents were north of 3500 for 2bd/2ba in most complexes (avalon, bayside, archstone). I just rented the same 2bd/2ba for 2700 (2800 with parking) with a better layout and more sq footage.

That’s a 23% drop (and some good shopping) for “somaboy,” and a tough trend in terms of (E)arnings for investors who paid a high (P)rice based on wildly different expectations.
Marcus & Millichap San Francisco Rental Outlook (And Quick Rebuttal) [SocketSite]
SocketSite’s Residential Real Estate Outlook For 2009 [SocketSite]

76 thoughts on “Marcus & Millichap Changes Their San Francisco 2009 Rental Outlook”
  1. So, M&M is now predicting 6% increase for the rest of the year. They should just wait till the end of the year to revise their prediction.
    [Editor’s Note: Not quite, previously calling for a 3.3% increase in 2009, now calling for a 8.9% decline.]

  2. People have been fleeing my complex recently. Every weekend there are trucks packing-up and moving-out.
    Vacancy rates must be through the roof as they have also reduced the asking rent on a 1-Bedroom from $2,400 to $1,850.
    I have a feeling that this has something to do with the 6-month severance checks coming to an end for many former Financial District workers that were laid off in Dec/Jan.

  3. I have a reasonably priced, IMO, rent-controlled 2 bd/1 bath w/ parking in the inner richmond neighborhood for $1,650 per month but have found that all of the prospective renters who viewed the unit barely qualify and are financially stretched.
    It is difficult to find good financially solid tenants.

  4. “[Editor’s Note: Not quite, previously calling for a 3.3% increase in 2009, now calling for a 8.9% decline.]”
    Asking/effective rent is already down 15% for the year. So the decline of 8.9% for 2009 means 6% increase for the rest of the year.

  5. Fish, do you have a link to a quantitative analysis showing a 15% drop? I wouldn’t mind showing that to my landlord…

  6. There have been more drastic decreases in SOMA. I wanted to move there last year but asking rents were north of 3500 for 2bd/2ba in most complexes (avalon, bayside, archstone). I just rented the same 2bd/2ba for 2700 (2800 with parking) with a better layout and more sq footage.
    Additionally I negotiated significantly with the mass of owned condos in the beacon, palms, 235 berry etc who are still listing at 3200-3600 for their 2bd/2ba but are very receptive to negotiated offers.
    Folks are a lot more price sensitive right now.

  7. “Fish, do you have a link to a quantitative analysis showing a 15% drop? I wouldn’t mind showing that to my landlord..”
    Try SFRentStats. Keep in mind, this is the asking rent advertised on the craigslist, not the current rent paid by all renters. The actual rent decrease that you can expect depends on the rent that you are paying now. For example, if you signed the lease at the peak of 2008 at the market rate, you could expect about 20% reduction if you were to sign for a comparable unit now.
    The best thing to do is to look for comparable units, negotiate the rent, and then take that to your landlord.

  8. Out in the suburbs, right next to the BART, 2bed/2bath units went from ~$1800/month to ~$1500/month over the past six months. Not that anyone would want to live in the ‘burbs. My point is, though, that the market is starting to react very quickly; he who hesitates (lowering the rent) is lost.
    My other anecdote is regarding commercial real estate. A VP at the company where my friend works is trying to find a larger building so that operations can be expanded. Problem is, every time they are ready to sign a lease, more space comes on line at a cheaper rate…

  9. It’s astonishing that anybody believed Marcus & Millichap’s forecast at all. Does anybody over there have a brain? Forecasting rental increases during what is the biggest economic crisis in 80 years?
    ROFL.
    Forecasting by “professionals” in America is a joke. I can’t tell if the forecasts are so bad because the intelligence level of forecasting employees is just that low, or is it because they are manipulating their data as part of some ill-advised PR campaign? (in other words, make cheery sounding forecasts at the request of the local landlords to see if they can bully prospective tenants into paying more in rent).
    I swear that I’m taking crazy pills, having to argue obvious points like “RE valuations and rental rates will fall during a huge economic crisis” and “SF will be affected by macroeconomic trends”.

  10. Spencer, depending on which building it is, that could be a good price.
    Anything lower than 2700 for a 2/2 is not gonna be a very quality product.
    I have my sights on a 2/2 in Soma/South Beach sometime in the next 3-6 months.

  11. Yeah, rents seem to be coming down a bit. The wife and I are working on locking in a new mortgage (4.5%, 30 yr fixed) for our building. The appraiser came out and looked things over. In his appraisal, he said that our unit (1450 sq ft, newly done kitchen, 2 br, 1 bath, 2 spots in the garage) would rent for $2k. We’re a couple of blocks from SF General.
    Now, I get it, rents are down, but c’mon, $2000 a month for that is freaking silly.

  12. Brian – Perhaps the appraiser was factoring vacancy into that $2000 number. The real rent would be higher, but when you amortize vacancy between tenants the effective rent becomes lower.

  13. I have a 1/1 unit that is turning over this month. It has super luxe, top of the line appliances and a gorgeous wall of windows opening out onto the garden.
    The current tenants asked for a 30% reduction based on craigslist ads they were seeing; other units really weren’t comparable when you consider the quality and newness. I get that some folks don’t care or want new or nice and just need cheap, but be careful what you’re comparing.
    I offered the tenants a 12% reduction, which they turned down, and got a new tenant after one showing with an 800 credit score. I suspect the old tenants were trying to stare me down into a greater price concession and have actually now asked for a few days’ extension past their notice until they get their new living situation sorted, which I can’t do.

  14. @Milkshake
    Yeah, I realize that, but it just smacked to me of “let’s find some lowball number, then figure out how to massage the numbers to arrive at that”. It’s annoying because even based on his (low) GRM, it’s dropping $100k – $150K off of the valuation. Yes, that’s a big number, and it also drops us below 20% equity (which means mortgage insurance, yuck).
    From looking at the link above, I’d say that we could get $2600 a month (easily) for our unit.

  15. So, what do you think about paying $3K for 2/1, ~1200 sqft, in pac heights?
    We signed in summer 2007 and the rent hasn’t changed. I seem to think it could be lowered, but average rents seem to indicate otherwise.

  16. Yeah, hard to say. I’m guessing that you got a pretty decent deal on your place. I guess I was going a step farther with the conversation than was maybe applicable (tying decreased rents to the decreased property values / appraised values).
    I think the problem is that, it’s too easy to just say “oh, rents are down” and then pluck a random number out of the blue that’s just not supported by anything other than “rents are down”.

  17. “The current tenants asked for a 30% reduction based on craigslist ads they were seeing; other units really weren’t comparable when you consider the quality and newness.”
    There are some clueless renters as well as clueless owners. I once had a guy asking for 2 mo free for a room that I advertised for $1500 in a $1.3m 2br unit. All they have to is to shop around for a few weeks to see a good deal when they see one.
    Make sure you have a deal on hand before giving 30 day notice, people. Offending your landlord is not good for relationship either.

  18. Brian – $2,000 for your place is crazy. I just rented out my 500 sq. foot Jr. 1 bedroom apartment in South Beach for $2,000 (furnished, full amenity building, 3 month lease).
    I had three people look at the place and all were earger to sign a lease. I probably could have gotten more, but I wanted to rent it quickly.

  19. Oh Milkshake, I can’t believe you posted that story. It’s enough to make me keep my two little rental units empty.

  20. Brian- unfortunately most appraisers are either clueless dumbasses or of the cover-your- ass variety. I can see why you’re infuriated, as that low ball rent/income number hits your loan value significantly. Guess I’ll be bending over soon as well, as I seek to refi a recently renovated triplex.
    Did you show the doofus appraiser rental comps or any data to get them to bump up their number? My god, if you have 2 garage parking spots, that’s $350-400 right there!

  21. “I just rented out my 500 sq. foot Jr. 1 bedroom apartment in South Beach for $2,000 (furnished, full amenity building, 3 month lease).”
    $4/sf is impressive, short term furnished or not. I wouldn’t pay more than $2200 for full 1 br (>750sf) w/ parking & view at Baycrest, Portside, Metro, etc. $2500 for Infinity and ORH.

  22. “I have a reasonably priced, IMO, rent-controlled 2 bd/1 bath w/ parking in the inner richmond neighborhood for $1,650 per month but have found that all of the prospective renters who viewed the unit barely qualify and are financially stretched.”
    Live Smart –
    Do you have a link to your property? I’m currently renting a unit in outer sunset for 2500 (rented peak of 2008 ugh) and would love to pay 1650 in inner richmond.

  23. Dakota…. I just rented out my 500 sq. foot Jr. 1 bedroom apartment in South Beach for $2,000 (furnished, full amenity building, 3 month lease).
    Good luck in 3 months – there are a lot of summer interns in the city May/June through August/September looking for short term rentals, once they head back to grad school you may be looking at a huge price reduction – maybe $1,200 a month!

  24. @45yo –
    No, we’re not supposed to talk to the appraiser, haha. So, we’ve brought up all of the factual issues with the appraisal, and I’m lining up comps as well. Oddly enough, he said that it was a very “complex” comp to factor, as there were very few recent comps. One of the comps he used was from July of last year! One of the rent comps he used is directly next to the central skyway (the elevated part of the freeway that dumps you onto market by Gough). Puh-lease. I’m still trying to figure out how that place has a higher GRM than my place.
    Anyway, I found a bunch of recent comps fairly easily (2721-2725 Harrison, 953-955 Valencia which is in escrow right now, and 25 Alvarado) – if I can find this stuff this easily, why can’t he?
    If I could rent this unit that I’m in for $2000 a month, I’d never, EVER consider [buying]. 2 parking spots, direct access to the back yard, and we have 600ish sq ft of storage space below us, too.
    Sorry if this is semi off-topic, just wanted to respond.

  25. 2 blocks from General Hospital? Which direction? That area doesn’t exactly command strong rents…

  26. which direction? toward Dolores park…
    Granted, I realize that I’m not going to be able to ask for $3000 a month for a 2/1 around here, but I do think that the market hasn’t dropped off so hard that $2000 would be the “market rate” for this unit.

  27. Brian- so you’re in the mish then (me too- live and have rentals here 🙂
    Here’s an idea- try to have them value the rent for the unit, and the rent for the garage seperately. Garages rent for ~$200 per car (easy to prove via craigslist), so that leaves $1600 for the unit (according to their logic.). A decent 2/1 in the mish w/o prkg is at least $1800, up to $2400 if very nice. Add +$200 per garage space.

  28. Brian- so you’re in the mish then (me too- live and have rentals here 🙂
    Here’s an idea- try to have them value the rent for the unit, and the rent for the garage seperately. Garages rent for ~$200 per car (easy to prove via craigslist), so that leaves $1600 for the unit (according to their logic.). A decent 2/1 in the mish w/o prkg is at least $1800, up to $2400 if very nice. Add +$200 per garage space.

  29. Live Smart
    I’d change the title to 1 bd + Den, 2 Car Outdoor pkg. If the minute I click on an ad, it’s disappointing, I can only imagine what’s going to happen when I show up. So I won’t waste my time.
    Then I’d add photos: 1 of the kitchen, one of the bath, one of the living room, one of the bedroom.
    Then I’d list the square footage. You can say approximately, but otherwise I think you need to come clean.
    Note that creditworthy tenants have much nicer places to choose from these days. 5 years ago, nearly every rental was a dump. Nowadays, that’s not true at all, there are a lot of very high quality places for rent in nearly every neighborhood. The best quality tenants are going to be attracted to the best quality rentals. The dumps are going to less credit worthy tenants.
    You can get a slightly better class of tenants if you allow pets. If the place is a dump, you’ll need to allow pets to have a prayer of finding a decent quality tenant and if it’s a dump, who cares.
    Finally, ads without photos are all dumps and no quality tenant even bothers with them.

  30. I agree with tipster. Personally I sort craigslist listings so that I’m only looking at ones with photos. I also would never consider living anywhere with wall-to-wall carpeting. (I’ve got a long list of maybe snootier demands — w&d in unit, at least 10 foot ceilings, top floor only, garage, etc.) but that one strikes me as pretty widely-held.)
    And I can’t stand listings that claim, e.g., 3 bedrooms, and then it’s clear from the text that it’s like one bedroom, a dining room, and a living room.
    Brian, $2k for a 2/1 that’s only 2 blocks from SF General strikes me as about right. Potrero over there is pretty ghetto and from what I recall the streets west of there aren’t tons better (and don’t get actually nice until after Valencia). My last 2/1 (with garage) was in a *much* nicer part of the Mission (really Dolores Heights) and I paid not much more than that (though it was admittedly a deal). Yours is at least a free-standing SFR though, right? (Mine was a flat.)

  31. I’m looking for a place (Russian, Telegraph hill, Pac Heights), and there are very few decent 1 bedrooms out there right now in my range ($2200-2800)(no parking, bay view). Any thoughts on whether more units will come out as the summer ends?

  32. @Shza:
    No, not SFR, 3 unit building. 1500 sq ft 2/1 (that’s really a door being moved away from legally being a 3/1), 2 parking spots, laundry in unit, newly remodeled kitchen w/dish washer, tons of storage space, etc. – and you think $2,000 a month is “about right”?
    OK, so explain listings such as:
    http://sfbay.craigslist.org/sfc/apa/1222581412.html
    (ROFL, walk through the second BR to get to the dining room) – $1850 without parking
    http://sfbay.craigslist.org/sfc/apa/1219452063.html
    Actually on Potrero facing SF General (much more noise from the road), no parking, $1950 a month – which pencils out to $2300 – $2400 a month if you factor in 2 parking spots…
    This place was rented out 7 years ago for $2200 a month – have rents really dropped *that much*?

  33. “I’m looking for a place (Russian, Telegraph hill, Pac Heights), and there are very few decent 1 bedrooms out there right now in my range ($2200-2800)(no parking, bay view). Any thoughts on whether more units will come out as the summer ends?”
    There was 2br/2ba unit on Union/Jones with spectacular view of Alcatraz/Bay for $2200 about a month ago. View alone was worth the price.
    If you can extend to SOMA, Paramount has one for less than $2200. (I’m paying $2100 for a spectacular view.) I’m sure you can negotiate This one for $2163/mo or less.

  34. This place was rented out 7 years ago for $2200 a month – have rents really dropped *that much*?
    That definitely proves that someone was a complete idiot 7 years ago. As I mentioned, I had little trouble finding a great Dolores Heights flat with 12 foot ceilings, garage, w/d in unit, and fantastic downtown views for $2150 in 2004. It was under-market then but I saw comparable places at the time for $2600.
    I’m currently paying $2250 for another 2/1 w/garage, w/d, high ceilings, bay views.
    I wouldn’t consider paying $2000 (or really anything) to live on the Potrero/Mission border but maybe that’s just me. One can certainly do better for the price.

  35. brian- you can get more than $2k for that trust me. If there is a way to make it a reasonable 3br, you could up the rent a few hundred more as well. As you know, plenty of hipsters here renting with roomates, so 3br = extra dinero, amigo 🙂
    As for the neighborhood, to each his/her own (no need for a marina vs. Misdion pissing match) but the neighborhood is in high demand due to aforementioned hipster factor (even for older, or sarcastic ones 😉

  36. Is *that* part of the mission really in demand though? (That’s an honest question. I find it incredible but I fully concede that you know better than I.) I can understand hipsters wanting to live at, say 23rd and Capp. But way out east where there’s nothing to do and you’re guaranteed to have random dudes hanging out and drinking on your front steps and then breaking into your car? Really? People want that?
    I loved living in “country club mission” where Delfina, Tartine, Bi-Rite, as well as tacos and everything on Valencia were nearby — plus I got to live on the “nice” side of the divide; but living blocks from the projects and even farther away from the good things in the mission seems like a lose-lose.
    It’s funny to see the “list your 2 BR somehow as a 3BR” thinking in action though. I can’t believe anyone falls for that — but apparently they do.

  37. Agreed that this is not “the mission”. Also agreed that 2K in rent will farther and in nicer hoods, too. Even 24th and bryant is nicer, and I have a friend renting a 2/1 there for about 1500. Maybe you wont get heloc-class finishes elsewhere, but I can’t see the market premium for a 13K fridge when you have to step over shit to get to the front door.
    I have friends who bought there, and last I checked were looking to leave (wife doesn’t feel safe there, kid is on the way, etc.) I think the plan was to rent the house out until the market improves. I once asked what they think they could rent it for: “Oh, about 5K” was the answer.

  38. cute article shza, but either that piece is over emphasizing the trustafarian factor in williamsburg, or the mish is different. sure, i had a few apply for units, but i got more googlers and other sw developers in comparison.
    as for what is the mish, the border is potrero/division/guerrero/cesear chavez. the ‘eastern’ part, east of so van ness let’s say, is loads of fun. great cafes on 24th st, and some awesome restaurants in the NEMIZ part of the hood (~ east of folsom, north of 20th st) like universal, slow club, flour+water, schmidt’s to name a few. i personally like it more than dealing with the bridge/tunnel crowd at tartine/bi rite/etc. and the neighborhood has true diversity with a huge variety of people. i feel safe here and walk around all the time, as does my wife. rents here are about 10% less than mission dolores area/baja noe valley.

  39. “over emphasizing the trustafarian factor ”
    Perhaps for those who are still purchasing units and homes, but not for renters. Maybe the Mission is still a neighborhood for people who actually work, but I have many neighbors and tenants here in the Marina, who spend hours doing everything but working. The Grove is full of sad empty Trustafarians who are recently not doing a very good job making their rent payments on time to me. A large percentage of my rental applicants are people who have income from family trusts or just “daddy”.

  40. Honest and sincere question:
    the last I rented was in 2004, when I was making $90K and I paid $1300 a month for my 1/1 apt. It felt painfully expensive to me.
    These days, everybody seems to be paying $2000 or more, how much do you all make? How much do you spend on everything else in life, like food, car, entertaining…….?
    Are you all CEOs?

  41. ester, I think the answer is that most renters in SF wouldn’t consider less than 25% of their take home pay to be “painfully expensive,” as you did. I know many people who pay at least 50% of their take home pay in rent, sometimes more.

  42. Well, the monthly rent was the single largest check that I wrote each month, by far. Isn’t that painful?
    If someone uses 50% towards the rent, what is the life quality like then?
    I would move away from SF, definitelyl.

  43. ester, rent is less than 10% of my take home, yet it’s still the largest single check that I write monthly, by far. I think that rent would be the largest single monthly expense for just about everyone except for the wealthiest of the wealthy, whose personal jet may cost more per month.
    If a person is bringing home $4000 a month take home and paying $2000 a month in rent, that still leaves a pretty healthy amount for their quality of life. If they don’t spend money on a car and aren’t saving much (common with SFers I’ve found), that’s a lot of money for food and playing.

  44. What’s the best way to ask a potential landlord for a lower rent?
    I have a great comp – an apartment was rented last week right next door, same view and size, built in the same era but rented for $50 less than this apartment and included all utilities and cable plus 1 car parking).
    I’d love to ask for a 5-10% reduction, but might still want the apartment even at the listed price. Will merely asking for a reduction harm my chances of getting passed over for another applicant?

  45. dub dub,
    Yeah, but the store pays only 20-30K/Year. These numbers don’t work for most BA workers. The business cannot probably survive with higher salaries which is why high-volume low-costs chains are slowly replacing mom-and-pop stores…

  46. “Will merely asking for a reduction harm my chances of getting passed over for another applicant?”
    What other applicant? Chances are, there is no other applicant at that price if the other unit went for less. Personally, I’d offer even less and see the counter. But then, I’m not in love.

  47. This is old, but if anything, the numbers are higher today:
    http://www.sfgov.org/site/moh_index.asp?id=5812
    This is the MOH report from 1999 on housing cost and affordability in San Francisco. Note how much higher the listed rents are here than from the census numbers reported by Robert.
    20% of all households spend 50% on rent, mostly lower income households, of course. As mentioned before, you can save on transportation costs if you live in The City. What I think communities should really measure is housing+transportation costs, to determine the total burden, especially for lower income people.
    http://www.b-sustainable.org/social-environment/percent-of-income-spent-on-housing-and-transportation-costs
    “Nationally, according to 2003 Bureau of Labor Statistics, for every dollar a working family saves on housing, it spends 77 cents more on increased transportation costs.”
    So moving further out doesn’t really save you that much, if you are working poor and your job in in San Francisco, but it adds to your commute.
    You can always get another job, bu the pay is much less, especially for low skill jobs.
    I also know lots of people, mostly recent grads, who seem to spend an inordinate amount on housing. I am not sure why they do it, I guess having your own place impresses the opposite sex or something. I think the largest housing burden I ever had here was when I first moved here and paid $400/mo for a room in communal housing and my income was $2500/mo.

  48. “Note how much higher the listed rents are here than from the census numbers reported by Robert”
    They are the rents for vacant/newly rented units, not rents paid by all renters in the city. They are close to asking rents on craigslist rather than census data.

  49. Um, Fishtarian, the numbers that Robert was using were only places rented in the past year (from census data), not for every unit. We were only looking at newly rented units in that discussion.

  50. NVJ,
    The data you cite is census data — I have the whole db on my laptop! The differences in the numbers presented there and the numbers that I’ve cited are that those are asking rents for units which failed to find a renter (vacant units).
    Vacant units, for obvious reasons, have “rents” higher than rented units. Of course, the rent of vacant units is zero, but the asking rents are higher.
    As you know, 1999 was a very tight year for rents. Don’t remember the vacancy rate offhand, but less than 1% seems right. Rents were through the roof. In that environment, vacant rentals are not only an extremely small sample, but by definition are the most expensive stock around, since they have failed to clear even in such an inflated market.
    In fact, trying to gauge the rental market by omitting all units which found a renter is a bit strange — no other market price, that I know of, is measured that way. House prices are tracked by sales, for example.
    It’s better to look at what actually gets rented (i.e. what renters pay), but restrict to recent renters. IMO this is a more accurate gauge of the market price.
    In terms of the renter burden, you are correct — as incomes go up, the proportion of income devoted to rent falls, so that, in 2000, the bottom 1/5 (those making 22K or less) paid 52% of income on rent. The top fifth (those making more than 100K) paid a median of 11% of their income on rent.
    If you restrict to recent renters, then the numbers for the bottom fifth and top fifth change to N/A and 14%. N/A because, while the median is 73% (!!), it is a very small sample size.

  51. Robert, just curious – how does the census obtain actual rents paid? Is it self-reporting by the landlords? By the tenants?

  52. All census data is survey data — so these are questionnaires. For specific information, a great resource is the iPUMS project, which has the text of questions asked, instructions to census takers, and a wealth of info about the methodology.
    The main advantage of using census data is the richness of the data. The decadal census has a long form sample size of 5% — really an amazing sample size. ACS surveys are 1% surveys, so often people combine three year surveys to get a 3% size.
    Alternately, you could try to mine IRS data and the like — the BEA does this, and combines it with census data. The IRS does not release micro-data, though 🙂
    One thing I’ve been thinking about is a way to more accurately track rents.
    It would have to be a craigslist crawl, in which you would track individual properties, identify re-listed properties, and only register the final asking price for each property, completely ignoring duplicates or “asking” prices until a property goes off the market. A price would be defined as “final”, if it is not re-listed, say within a month. This means all my data would be 1 month old. If further price adjustments occurred between the tenant and owner, I would not see it, but the result would be far more accurate than the current web crawl approach. I would also get data such a mean time-to-sell, and possibly even frequency of re-listings (i.e. is this a 3 month rental) which would allow for calculating effective rents, etc.
    I’ve been looking into this seriously: for example, I took a 10% sample of asking prices for 2 bedrooms in the 1000-1500 price range, and again 10% sample in the 1500-2000 price range. By hand, I went through every listing, and counted how many times it was re-listed. The “higher” income bracket listings appeared an average of 3.2 times within a 1 week period. The lower income listings were re-listed only 1.2 times, on average, in a 1 week period. So, if you were to write a “dumb” spider that would just collect all the listings and compute medians, there would be a substantial skew because of duplicates.
    Thinking about these issues really brings to light the trade-offs when trying to get good data: timeliness vs. accuracy, model data vs. sample data, etc. For example, there would be seasonality (school year moveins) so I could calculate my own “birth-death” model that would adjust for this, but then this would miss turning points — it really is an interesting project, but craigslist frowns on spiders, so I’m not sure if it would work.
    If it does, I will post my findings here!

  53. he differences in the numbers presented there and the numbers that I’ve cited are that those are asking rents for units which failed to find a renter (vacant units).
    I don’t think that is true. MOH says that their source is RealFacts Inc. Looking at their website, they sell market comp data, presumably mostly for landlord use. I don’t know their methodology, but they must do more than just look at asking rents.
    I think that prices are actually a little bit higher now than in 1999. What I can find online supports that. It is pretty close, in any case.

  54. Well, the data seems to be the same as that coming from the census summary tape file 3, for the variable “rent asked (for vacant unit)”, controlled for by the number of bedrooms. So, at least in this case, RealFacts is not adding much more value than a db query on census data. I think you could fiddle with the online factfinder site to verify this, but the db access there is clumsy.
    To the best of my knowledge, only the census does large-scale surveys such as this. The other data I’ve seen consist of either sampling classified ads in newspapers (HUD/OFHEO have done this) or looking at a craigslist crawl, which is basically the same as the above, with some distortions due to zero publication costs.
    I would be interested in other data, though, if you can find it.
    On the other hand, perhaps the methodology is different and they are coming to the same conclusion.
    I do agree that rents are higher “now” than in 2000 — about 25% higher, measuring 2000-2007. As to what happened recently, I don’t have that data, and it seems hard to come by — funny how these reports match census survey years 🙂

  55. “Um, Fishtarian, the numbers that Robert was using were only places rented in the past year (from census data), not for every unit. We were only looking at newly rented units in that discussion”
    Hm, ok, nobody corrected me on that thread. Then my conclusion still would be that craigslist overstates the magnitude of (real market rate), but a useful measure of changes (in asking rents) over time.
    I don’t think it matters much if over-priced properties get listed longer, unless you insists on complete accuracy in magnitude. The skewing in changes is not significant enough as 20% drop in asking attests. In any event, they are what the market is asking for, over-priced or not, and one just needs to bear in mind that the asking price is softer in soft markets.

  56. I don’t think it matters much if over-priced properties get listed longer, unless you insists on complete accuracy in magnitude.
    It matters a great deal. To be frank, I was taken aback at the resistance I was getting on that point, but with some hindsight, I understand now.
    It reminds me of when I was looking at the 1930 rental data, and I saw that the median asking prices for vacant units was much higher than the median prices of rented units — obviously not because of rent control, or increasing prices, as there was a raging deflation. Then I looked at the 1940 data and the 1950 data — same story. NVJ just posted the same data for 2000, btw.
    Regardless, the reason for the discrepancy is exactly because some units are listed longer.
    You see, when you want to measure the market price, you are interested in transactions — not advertisements, or offers of transactions. The median “offer price” represents the supply curve, but there is also a demand curve, the quantity of units transacted is when the two curves meet. So, if a butcher offers meat at $10, and then again at $8 and finally a sale at $6, the median price would be $8. But, if very few people buy the $8 meat, and more buy the $6 meat, then the median transaction price might well be $6. So, you cannot just measure supply — you need to measure demand as well.
    Now, you can argue, “apartments aren’t meat!” and it’s true: each apartment is unique. But, each apartment doesn’t take the same amount of time to sell. When prices are high, they are high because there is a shortage of free apartments. Almost by definition, this means that fewer transactions are occurring at the higher price. If there is a glut of inventory, the price drops to the point where more demand is drawn out, and more people choose to rent in the city. So, the median of the transactions will always be lower than the median of asking prices, particularly over, say, a 2 year period. More transactions occur when asking prices are low. So, the “average” person will have rented in a time period when rents are below average.
    In addition to this, you have differing market segments. More expensive products appeal to a smaller market, and must be advertised more. If it takes a higher priced apartment 2 months to sell, and a lower priced apartment sells in 2 weeks, then the higher priced apartment will have 4 times the influence on the asking prices, but should have exactly the same influence on the transactions.
    Finally, even within the same market segment, apartments “above” the fair market price can linger for months, being re-listed every 2 days, whereas those that are below disappear within a week.
    Because of all of these effects, you can easily get into a situation in which the median transaction price is substantially lower than the median Cragslist asking price. So, it matters a great deal.
    Btw, this has nothing to do with apartments. It’s true in most areas. For example, just by counting the asking prices of car commercials, you would assume that the median car sold would be a higher end lexus or prius, for example, when of course they just advertise more — they don’t execute more transactions. The luxury market is irrelevant to medians of car purchases — this is obvious of cars, houses, and most things, but not, supposedly, for apartments.

  57. Robert, I think the main reason that you got a lot of pushback was that the medians you listed from census data are lower than what any of us ever saw for any apartment of those sizes during those years. For me, it means that A) something is weird with the data, B) I’m really bad at negotiating apartment prices, or C) there are a lot more apartments that go rented each year that I had no way of finding or seeing when I was looking for one.
    It’s probably a combo of B and C, but I still wonder about A. I mean, I wasn’t picky about where to live when I first moved here in 2002, yet I never saw anything in the price range (aside from a couple in-law places with 7 foot ceilings) near the median that you listed, and that means that somehow I missed more than 50% (!!!!) of the units out there! Just seems weird.

  58. “Because of all of these effects, you can easily get into a situation in which the median transaction price is substantially lower than the median Cragslist asking price. So, it matters a great deal.”
    Yes, I already conceded that point. My point, however, is that it’s the *trend* that matters to a buyer like me, not the median asking/transaction price, which I have absolutely no use for. For example, let’s say median asking in SOMA dropped from $3000 to $2500. Using that data, I can expect certain reduction in rent when I resign. Or price a Palms unit that went for $3400 a year ago. But I can’t use the median price to price any apartment unless I know that it is of median quality.
    Now, if you can make an argument that the trend captured by SFRentStats is skewed, that would be of more interest to me.

  59. Fishtarian, yes, I think I have made this argument, and in fact you can check for yourself, by going on craigslist, and computing the medians directly, if only you do not count re-listings more than once.
    Furthermore, if you keep track of the listings, and only “count” them once they are removed (i.e. sold), you will get an even lower number.
    Now, automation would be best for this approach, but you could do a sample by hand. That should be easy to check and may help you. Good luck.

  60. I think the main reason that you got a lot of pushback was that the medians you listed from census data are lower than what any of us ever saw for any apartment of those sizes during those years.
    Well Anon, I certainly understand your point of view, but I think this fits into the whole delusion thing — San Francisco has a lot of middle class and below people — and a ton of not-so-nice housing stock. I don’t recall the self-reported incomes on this site, but they are an outlier. I also hear talk of hundreds of thousands of dollars in income being “middle class” and hidden wealth, rich foreign buyers, etc.
    So self-perceptions differ from reality — by a lot. It would make sense that apartments/locations that you would not even consider to be in your universe may alter medians quite a bit, and this might explain the perceptions.
    For my part, I know a lot of grad students/post-docs/waiters and the like, and the figures I cited matched my experience in terms of what they paid. None of them lived in dangerous slums, by the way, but I guess most of the posters on this board would never consider living in those types of quarters.
    Here is one way to do a gut check on these things: if you are a landlord, and you think you are renting out an “average” apartment in terms of cost, then look at the finances of your tenants. Let’s drop the bottom 15%, since we assume they are on public housing. Even then, the median income is around 74K — say 70-80K. If your tenants consistently earn more than this, then you are not renting out an “average” apartment. Maybe the location is better, the size, or the finishings, etc, but in either case your sense is not that of the median.
    If your tenants have household incomes in excess of 100K — then they are in the top third, in our non-subsidized renter universe. So basically whatever agreement you come to with them about rent is irrelevant for medians.
    By the way, you can definitely get higher rental data out of the census: the median rent paid for those earning 100K or more in 2007 was $2337, for recent occupants. So, say mid 2006-mid 2007. I can definitely see how, if that is your universe, then that’s what you would consider normal. In fact, I bet if you polled the people on this site, and asked them what they actually paid, that this “micro-data” would not be much different than the available census data.
    I guess I am on a one-man crusade to point out that most of the housing is not as nice as is featured here, and most of the incomes are much lower than people believe, and that in general there is a gap between perception and reality. In my experience, people would rather believe that there is a conspiracy on the part of the census than to think that they are living in a privileged bubble. There are a lot of interesting social/emergent reasons for this.

  61. “Fishtarian, yes, I think I have made this argument, and in fact you can check for yourself, by going on craigslist, and computing the medians directly, if only you do not count re-listings more than once.”
    Robert, I’m talking about the changes in median, not the median itself. I already agreed that the craigslist median is not reliable. Now I’m asking you or someone to show me that the changes in craigslist median is also unreliable.
    If median is consistently biased, then the change in median is still useful. Eg, if craigslist overstated the median by 15% both in 2008 and 2009, then the drop of 20% in craigslist median is a reliable proxy for the market drop.

  62. “I already agreed that the craigslist median is not reliable.”
    Actually, for the record, I shouldn’t diss craigslist this way. There shouldn’t be an issue of reliability as long as one understands that craigslist merely represents the asking price, not transaction price: it’s still the asking price, no matter how many times over-priced properties are relisted. And attempt to derive transaction median by processing craigslist data probably isn’t going to work since the actual transaction prices are not published on craigslist: the final listing price is not necessarily the transaction price.

  63. Well, it’s not a question of things “not working” — the question is, would properly handling repostings be more accurate than data as published on SFrentstats? Would counting only the “terminal” listings further improve the accuracy? I think most would agree that the answer is yes.
    Moreover, such an approach would yield information about *volume*, which is critical, and might clear up a lot of confusion about the market.
    For example, a fall in SFRentStats prices might be because the more expensive apartments are selling more quickly, not necessarily because prices are falling. So, you need data on volume to interpret both the actual price levels and the changes in price levels.
    As you said, craigslist is the market, and all the info is there. It’s just unfortunate that people depend too much on a site such as SFRentStats to interpret that data for them.

  64. Robert, I hear what you’re saying regarding posters on this site and the apartments that we’re looking at. The reason that I singled out 2002 is that I was a grad student at the time looking for the cheapest place possible that I wouldn’t bang my head on the ceiling. I ended up splitting a three bedroom place in the Inner Sunset with two other grad students, for FAR above the median numbers that you posted.
    If I could have found anything near the median numbers anywhere in the Sunset, Ingleside, Richmond, etc, I would have been able to get a one bedroom pretty easily. (I wasn’t looking in prime areas, I wasn’t picky, etc – I was a broke 24 year old grad student) I didn’t find that to be the case, which leads me to believe that there must be a TON of listings that are rented outside of the normal channels OR I’m simply terrible at negotiating prices.

  65. Hey Anon, well, to be clear, I didn’t list any medians for 2002. There is a quirk in the micro-data source I have so that it only gives the decadal census data (2000), and the ACS are available only for 2005-2007.
    Btw, I also rented my current place when you did, for 1350/month. A 2/1 in Noe — outer Noe would be more correct, near the nunnery. Bottom unit, but I think it’s a nice place. Tall ceilings, but dingy split bathroom. Not that I am one to care about bathrooms or kitchens. It was about three years before I realized that my kitchen has granite on the countertop — I didn’t notice the surface until everyone started talking about granite countertops, which I thought was a very strange thing to be talking about. Out of curiosity, I went into my kitchen and looked. Lo and behold, some clumsily placed granite was there. It stains if you leave a wet object on it.
    But, I distinctly remember that it was a buyer’s market. At the time, I didn’t know anything — a grad student friend of mine wanted to move to the city with me, and I agreed, with the proviso that we not pay more than what I paid in mountain view. So we looked up 15 apartments on craigslist, all for about 1400 a month, and the landlords were enthusiastic. I kept forgetting which apartments looked like what, and the addresses didn’t mean much to me, so it was confusing trying to decide which one to take. The property manager at the place we finally rented offered to knock 50 bucks off the rent if we would take it, and we did — so my negotiating skills were poor then, too. I’m sure I could’ve gotten at least *one* of those 15 places for 1200. To give you an idea of how ignorant we were, my friend was doing the paperwork, but one day he needed to be on campus, so he gave me the notebook with all the phone numbers, and asked me to schedule a few more appointments. I called the first number, and the lady asked which unit we were interested in seeing. I realized then, that I was unprepared, so I asked her “which apartments do you have?” and we went in a circle for a while. Then, I remembered that we wanted to be in “The Mission”, and I recalled some street signs, so I asked her “Do you have one near 29th and Valencia?”
    I was just in Hayes Valley tonight, and bumped into a girl I met earlier. At that time she told me she was looking for a new place, and I advised her to negotiate. Today, she told me she did, and got $150 reduction. She rented a place on Laguna and Fell for 1650 — a 2/1. Her place, though, is not on the top floor, and she said it was “quirky”. I didn’t press for details, but the location is nice. She is also a foreigner, and not too savvy — taking rental advice from me — but she did fine.
    I also met a friend of mine there (Momi Toby’s) and she pays low rent — around $1000, for a 1 bedroom. It is *huge* with very high ceilings, but an awkward layout — you need to go through the bedroom to get to the kitchen. However, this place is seriously water damaged, and she got it via word of mouth from the landlord, who really does not put any money into it. I’m sure it violates all sorts of codes. She is very handy, if that is the appropriate term for a professional welder/product designer/precision machinist — and ended up fixing some stuff, but dealing with rot in the lathe and plaster walls is hard to do. She also described various incompetencies that the landlord did in improperly trying to cover up a fireplace with plaster that wouldn’t adhere to the brick, and poorly installed fixtures, including a leaky skylight, all of which she fixed.
    The vast majority of the housing stock in this city is decaying due to water-damage, which is one side-effect of high house prices: owners think they will always be able to sell for more, so why bother with dry rot, or fixing leaks? Moreover, they tend to be under-capitalized, which is what you would expect in a town filled with many small time operators, most of whom are in way over their heads, in terms of being able to preserve antique victorians.
    So, the long and short of it is that behind that pretty facade is a lot of rot. We are not that rich. Rents are not as high as many believe. The houses are rotting, and the “rich” landlords don’t have enough capital to maintain them, neither do they have as much equity as they believe, neither is San Francisco the “center” or “core” of anything other than entertainment. This is a fun, cosmopolitan city, but in order to keep on having fun, we hold our noses and avert our eyes, eventually believing lies about where we are.

  66. oops, I want to admit to an error, which is that the decadal census data is always for the prior year. So, we are conducting the 2010 census now, and the data I cited for the 2000 census measured rents in 1999, which means that recent movers would include those moving in from 1998-1999 — this would be a lower median than for those moving in from 1999-2000, due to the raging IT bubble.
    ACS data *is* conducted in the cited year — Although the months differ depending on the survey and city. Don’t have that at my fingertips. Sloppiness on my part — all census year data I cite needs to be viewed this way, except 2005-2007 ACS data.

  67. “It’s just unfortunate that people depend too much on a site such as SFRentStats to interpret that data for them.”
    Not sure what you mean by “depend on too much”. Though SFRentStats is not transaction stats, we know from RE market that the changes in asking price median tracks the changes in closing price median, roughly or closely, depending on your precision. (Overpriced rentals getting listed longer is really no different than overpriced properties having longer DOM, is it?) We just need to keep in mind that it’s an asking price stat, not transaction stat, and that there is more room for negotiation in down market.

Leave a Reply

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