The number of single-family homes and condos that traded hands across the greater Bay Area totaled 6,825 in December, down 5.9 percent from a downwardly revised 7,250 sales in November and 3.9 percent lower versus the same time last year.
And in San Francisco, the recorded sales volume dropped to 427 last month, which was 17.7 percent lower than in November, 21.1 percent lower on a year-over-year basis and the slowest December since 2011, according to recorded sales data from CoreLogic as charted above.
In Alameda County, recorded homes sales slipped 1.2 percent from November (1,556) to December (1,537) and 1.1 percent lower versus the same time last year and sales in Contra Costa County dropped 3.5 percent to 1,360, which is 5.4 percent lower on a year-over-year basis, while sales in Solano County actually gained 10.5 percent in December to 612, which was 3.6 percent higher versus December 2016.
Home sales in Santa Clara County slipped 1.9 percent from November (1,550) to December (1,520) but and were 5.2 percent lower versus the same time last year while recorded sales in San Mateo County dropped 11.8 percent from November (621) to December (548) but were effectively unchanged (up 0.2 percent) versus 2016.
And having taken a fire related hit in October, home sales in Napa ticked up 6.6 percent from November (121) to December (129) and were 4.0 percent higher on a year-over-year basis while sales in Sonoma County dropped 21.1 percent from November (597) to December (471) but were 1.7 percent higher versus the same time last year, and sales in Marin dropped 31.6 percent from 323 in November to 221 in December and were 9.4 percent lower versus December 2016.
The median price paid for those aforementioned 427 homes in San Francisco was $1,174,000, down 9.7 percent from the record tying median price the month before but 6.7 percent higher than at the end of 2016.
The median sale price in Alameda County was $760,000 in December, down 3.2 percent from November but 12.6 percent higher versus the same time last year; the median sale price in Contra Costa County was $550,000, down 5.4 percent on a year-over-year basis; and the median sale price in Solano County was $410,000, up 1,2 percent from November and 5.1 percent higher, year-over-year.
The median sale price in Santa Clara County was a record $1,000,000 in December, up 2.0 percent versus November and 24.2 percent higher versus the same time last year while the median sale price in San Mateo County was $1,200,500 in December, down 5.8 percent from November but 19.1 percent higher, year-over-year.
The median sale price in Marin was $950,000 last month having slipped 0.2 percent from November but was still 10.8 percent higher versus the same time last year while the median in Napa was $632,500, down 2.7 percent from November but 12.9 percent higher versus the same time last year, and the median in Sonoma was $615,000 last month, unchanged from November but 16.9 percent higher versus December 2016.
Across the greater Bay Area, the median home sale price dropped 4.5 percent in December to $750,000 but remains 12.1 percent higher versus the same time last year.
Keep in mind that while movements in the median sale price are a great measure of what’s selling, they’re not necessarily a great measure of appreciation or changes in value and are susceptible to changes in mix, as opposed to movements in the Case-Shiller Index.
– “The median price paid for those aforementioned 427 homes in San Francisco was $1,174,000, down 9.7 percent from the record tying median price the month before but 6.7 percent higher than at the end of 2016.”
– “The median sale price in Santa Clara County was a record $1,000,000 in December… 24.2 percent higher versus the same time last year while the median sale price in San Mateo County … 19.1 percent higher, year-over-year.”
– “The median sale price in Marin was $950,000 last month … was still 10.8 percent higher versus the same time last year while the median in Napa was $632,500.. 12.9 percent higher versus the same time last year, and the median in Sonoma was $615,000 last month, unchanged from November but 16.9 percent higher versus December 2016.”
That really highlights what’s happening in the bay area…. We’ll see how the borrowing rates and the adjustments to tax laws affects things going forwards but we can all agree that it has a chance of continuing..whether we like it or not
A couple of other key paragraphs you seem to be missing missing:
“[In] San Francisco, the recorded sales volume dropped to 427 last month, which was 17.7 percent lower than in November, 21.1 percent lower on a year-over-year basis and the slowest December since 2011.”
And of course:
“Keep in mind that while movements in the median sale price are a great measure of what’s selling, they’re not necessarily a great measure of appreciation or changes in value and are susceptible to changes in mix, as opposed to movements in the Case-Shiller Index.”
Low inventory = low sales volumes. And median prices up 6.7% YOY. Again, unless and until supply increases or demand wanes, or both, we’re not going to see any meaningful slowdown or reversal in pricing. Neither has occurred yet.
Listed inventory levels in San Francisco actually bottomed out in 2015. In fact, the average inventory level was 9 percent lower in November/December 2015 versus 2017 while December sales were somehow 12.5 percent higher in 2015 as compared to last month.
And then ticked up in 2016 (albeit still at low levels) and then fell again in 2017 and remain at very low levels. And the inventory of houses was lower in 2017 than 2015 (condos higher). That extremely low supply in 2017 was even lower in 2015 when the market was white hot doesn’t really say much. Low supply = low sales volume and high prices given the continuing high demand.
And inventory levels in San Francisco were, on average, over 20 percent lower in November/December 2014 versus 2017 while December 2014 sales were 10 percent higher than last month. So perhaps “low inventory = low sales” (except when it doesn’t) isn’t the best way to explain away the data and trend at hand.
You’re comparing very low levels of inventory (2017) to extremely low levels (2014). As I noted, an argument that “it’s not quite as strong as 2014 when the market was absolutely on fire” is not particularly compelling.
Got it. So the market was simply stronger in 2014/2015, resulting in higher demand despite lower inventory levels at the time.
Investors = GONE. Mostly to Seattle.
2nd Home Buyers From the Peninsula = GONE, thanks to higher taxes. Cheaper to rent a hotel room.
Renters seeing ever higher rents and buying to lock in a housing price = GONE, thanks to lower rents and higher taxes.
I’m not seeing the driver for any increase in 2018-2019. In fact, many of the people who bought pieds-a-terre in SF will likely sell them now that they are costing an extra $500-1000/month in non deductible property taxes, when they only use them a couple of times per year. You’ll see this en masse next year when people go to file their taxes and get socked with a huge bill from the non deductibility of their property taxes.
Investors are mostly gone to other markets in addition to Seattle. Southern Washington/Portland, Reno to name two others on the West Coast. The tax bill will have an effect and increasing mortgage rates but I still see SF appreciating at the national average rate for the coming 5 plus years. So I’d guess there will be an increase in 20018-2019 in SF. Very modest and excluding possibly recently built condos.
But people who live in San Francisco and earn a high salary are not gone (sorry, NOT GONE).
They’ll still need a place to live and CCSF will continue to limit the supply of new housing.
Now City College is responsible for the housing bubble? Ben Bernanke will be relieved to know his trillions of $$ of quantitative easing and years of ZIRP aren’t the culprits!
“City and County of San Francisco”
Once those people see their tax bill a year from now, people will start seriously question what the hell they need to be here for, and the masses will start drifting away.
“Once those people see their tax bill a year from now, people will start seriously question what the hell they need to be here for, and the masses will start drifting away.”
Hate to break it to you, but the millionares are still getting a tax break even with the tax changes. I work with high net worth individuals, most were getting their deductions limited by the AMT in the past. Reduction in the rates, and changes to pass through entities still benefit them.
Looking at the stock market over the past few days, I’d say the tech companies might have seen peak growth for the time being. I’m not saying they are going away by any means, but the low hanging fruit is gone, it will be more competitive moving forward.
Facebook had its first month without North American user growth, as one example. New advancements in the phone/app space are probably going to be fewer and far between, while the next big thing of robotics/AI is not ready for prime time. Not great news for Bay Area real estate, with already nosebleed prices and lots of new construction.
The robust tech growth is maturing to an extent but not ending. Slower growth maybe but still growth. It’s not clear how AI/robotics will impact the BA. I don’t see the BA as being the center of these industries as it was of tech. Prices will continue up but at a slower pace.
I think the social media type of techs (w/o any other problem solving purpose) are over. Smartphones have had its run. But there are still stubborn areas (healthcare, government, courts/law) which have room for technological advances to enable them to deliver excellent services at minimal costs.
The adage, “find a solution to an existing problem and you will be handsomely rewarded for it,” will always apply.
If investors = GONE, how come investment props in SF are still selling like hot cakes, and still selling near top values?
“2nd Home Buyers From the Peninsula = GONE, thanks to higher taxes. Cheaper to rent a hotel room.”
No matter how many times I tell one of my clients that it is cheaper to rent a hotel room and she should sell her second home, she refuses to do it even though she probably spends less than two weeks a year in it. Some 2nd home buyers don’t actually do the same financial calculation that you do Tipster.
I agree that the market is softening, but do not think you are going to see a mass of people selling because of the changes to the tax law. While it may have a marginal effect on demand, I do not agree with your prediction that we will see it happen “en masse”.
Tipster’s been spouting nonsense on this site for years. All we need for some real nostalgia is to bring back the the old Fluj/LMRIM debates.
SF, Santa Clara, San Mateo, and Marin counties will be hurt by the new property tax and interest deductibility tax laws. Oakland, San Leandro, Hayward along with most of Contra Costa County will benefit from the new tax laws. Places like Alamo, Danville, Lafayette, Orinda, Moraga in Contra Costa County will also see sales weaken. Oakland’s Eastlake, San Antonio, Fruitvale, Laurel, Maxwell Park, Milllsmont neighborhoods should be the big winners.
It’s not going to be as much of a factor as you think. Remember the 10K cap applies to state income tax payments too. Many of this group are making 150K or more and paying 6, 7 or 8K and more in state income tax. That leaves very little left over for mortgage deductibility. The big winner will be Washington State and Nevada to the detriment of RE in the whole Bay Area (southern Cali too).
New Marin homeowner here. Yes, the deduction cap on mortgages and limitation on SALT effect me but to be honest, those are offset by the income tax reductions. My realtor said lots of Marin home buyers aren’t even flinching because it balanced out for them too. At least in Marin, the drop is because they’re so few homes on the market.
Dave, the mortgage interest deduction does not factor into the $10,000 state tax deduction cap. Two different animals.
Sorry, I meant property taxes. Same argument holds insofar as I don’t see Oakland gaining a huge advantage here over SF or SM.
Hey, we agree on something!
I agree. I think a lot of people here are overestimating exactly how much an impact the new tax laws are going to have. It will have an effect at the margin and will be a slight drag on price levels, but it isn’t going to cause enough people to sell and move to really drive the market.
State of the overall economy, stock market, employment, and interest rates are all going to be more important factors in 2018 and beyond.
Taxes are the tail while overall economy, stock market, business sentiment are/is the dog. We know interest rates will go up simply because of the inflation which comes with high employment numbers. Technology and efficiency will lay many businesses to waste and the remaining players will capture even greater market share.
The original House Bill was much worse.
Sales price goes up – sales volume goes down. Extend these trends to their ad finitum and in 2035 exactly one house will be sold here for $100,000,000. It’s unsustainable.
The sale price has just gone up in SF, SM, SC and Marin due to the new tax laws. Therefore we will see fewer sales. The home prices will have to come down to move a high volume of properties. The Republican tax plan just took away equity from these homeowners.
Up 19% y-o-y- in San Mateo County! Wow! This just gets better and better! I just got on Zillow and ran all the numbers and sure enough, on paper, I am up even more than I realized.
We heard the: “Increasing Tax rates” and “Tech is maturing” arguments above. Let’s look at the jobs/income argument for a second:
– Unemployment is at record lows
– SF Commercial Real Estate recorded the 4th best year in history. Some of the largest leases in 2017 were:
SQF Company Quarter
736,000 Dropbox Q4
404,000 Facebook Q3
225,490 Slack Q1
207,000 Okta Q4
176,000 Amazon Q3
166,460 Google Q1
131,000 Otto (Uber) Q2
127,000 Atlassian Q4
107,000 Wish Q3
104,667 Adobe Q1
100,630 Accenture Q1
84,000 We Work Q3
83,000 Airbnb Q4
76,000 Affirm Q2
62,000 Lyft Q3
58,000 Postmates Q4
57,000 Cloudera Q2
56,000 Anaplan Q4
54,000 LendingHome Q3
None of us can look into the future, but maybe it’s worth asking the following questions:
– Are these companies (who actually put their money where their mouth is) expanding or shrinking their SF workforce?
– Are they likely to pay well above-averge wages?
– Where will the new workforce live?
And yet there were 4 straight quarters of net negative office absorption last year. The first time that has happened in about 10 years. Some of the leases may have been anticipatory of the Prop M cap being hit with relatively small allotments available going forward. Part of the boom in office construction was due to a backlog of M allocation which effectively allowed more than 850K of office space to be approved each year.
Dave,
There are 3 arguments why the slightly negative absorption rate is not necessarily bad news for residential real estate in SF.
1) While new office space supply is slightly outpacing demand, both are (and have been for years) rising at a very fast pace. More occupied office space means more jobs means more workers in SF means more renters and/or homebuyers.
2) Rents per SQF are still modestly rising. They have been doing so for the full 2017 year. While this trend might not continue endlessly, it underlines how strong the demand is in SF.
3) You have been singing your ‘swap Bay Area for ‘Seattle’ song since 2014. Every fact gets convoluted into something that fits your desired narrative. If you made such a killer move back in 2014, why then the constant Bay Area bashing and need for reassurance?
Keep in mind that the overall increase in asking rents for office space in San Francisco was less than the rate of inflation last year.
In addition, the average asking rent for older Class B space actually dropped 2.8 percent in 2017 while its vacancy rate jumped from 6.3 to 8.6 percent.
In other words: Development [has been] Driving the Office Market in San Francisco.
Good points. I see the market in the following way:
The Facebooks and Dropboxes (Class A space) are pushing out the Delta Dentals and Blue Shields (Class B space). Jobs paying $175k per year are pushing out jobs paying $75k per year.
People making $75k per year were anyways priced out of the market for SF SFH, while people making $175 are not.
For many tech companies, class B space is not really an option as they aim to attract the best and the brightest.
Btw, I’m not defending those trends, but this has been the reality in SF for a long time now.
People making $175k/yr are definitely priced out.
People making $175k are priced out of the most expensive homes, but they’re definitely not priced out of San Francisco.
As to Class B space, even though the under 50K M cap has available allocations, might it not be less expensive and time-consuming to convert a Class B building to Class A rather than build a new 49K building?
The other things about converting office space is, if you have an existing 100K Class B building and convert it to Class A, the developer then has 50,001K more Class A space than the new project built under the 50K M cap. Maybe this is going on but one doesn’t hear about it and it’d seem, with Class B vacancy rising significantly and rents falling, a way for an owner to effectively exceed to 50K cap and get “new” Class A space. .
The net negative absorption for the year is the first in a decade. While Class A space filled up, Class B space emptied out – and not necessarily B tenants moving to A space. Tenant mix is different with many B tenants unable to afford A space. More occupied office space means more jobs (short of companies banking space which may be a factor given M really kicking in now), but the negative absorption and big jump in Class B vacancy would suggest tenants – Class B tenants – left the city.
It has nothing to do with Seattle though following the crane contest has been fun. My investments there are doing great. It’s not BA bashing, its an honest critique based on observations (will there be negative absorption this quarter and have 5 quarters of negative absorption occurred before) from someone who has lived almost all my life here.
Again, there was much more occupied office space in SF at the end of 2017 than the year before. This is what counts for jobs and the demand for housing in SF. Net absorption rate is really secondary.
That’s factually incorrect.
Granted, it was only a difference of 95,000 square feet, but total occupied office space in San Francisco, not including vacant space available for sublet, was actually higher at the end of 2016 (71.6 million square feet) than at the end of last year (71.5 million).
So applying the correct data to your stated model (trend in occupied office space = trend in demand for housing), it would suggest that demand for housing is actually down in San Francisco, not up.
That’s very interesting. In 2017 there must have been hundreds of thousands of new office square footage delivered to the market. How come the total occupied square footage shrank from one year to the next? Is all the new office space (such as the Salesforce tower) remaining vacant?
@pero: Say they added 1 million square to 10 million at the end of 2016 (using “made up numbers” to illustrate) bringing the total to 11 million in 12/17. Say all the new space was leased (new space not sitting empty) but 1,000,095 million of existing space (Class B and Class A) was vacated. At the end of 2017 total occupied space is 9,999,905 feet – less than the 10 million at the end of 2016 – while total space is 1 million more feet than at the end of 2016. .
Dave, great example. But here is the but:
If 1m new class A space simply pushed out 1m of class B space, then we should see a significant jump in vacancy rates for class B (1m/11m = an increase of 9 percentage points). We saw a jump of vacancy rates for class B office space from 6.3% to 8.6%. If we know the total class be Market in SF, we can see if this supports your thesis.
Pero – yes, we need to know the mix of Class A to Class B and factor in the large increase in Class B vacancy to those numbers. There is some correlation between Class A and Class B space but, IMO, not a big relationship. New Class A space opening up does not get automatically filled by Class B space being vacated. Many Class B tenants can’t afford Class A space. My CPA’s firm has 8 partners and about 25 employees. They vacated their Class B space this year for space on the Peninsula as the parking was getting to be too much of a problem for clients.
The net absorption chart reinforces that, even as Dropbox and Facebook and Stack and others signed new leases during the year, that added space was more than offset by space being concurrently vacated.
Actually another factor in the office space class warfare: all classes of office space shot up in the last few years. Class A still did well in 2017 because top tech co’s wanted the prime space. Class B tenants had more options, hence the vacancy is now greater. Class B landlords were/are willing to keep vacancies longer to get the price they want. This is similar to neighborhood retail vacancies in the Castro, Valencia, etc. The detente will end with either landlords dropping prices, or other/new tenants signing up for class B space. Depends on numerous business and economic factors as to which way this will sway…
Given the significant drop in rents for Class B space along with the significant jump in vacancy, it appears Class B commercial landlords are not holding and hoping. Unlike, seemingly, retail landlords. Class B tenants tend to be smaller and have more flexibility. My CPA’s firm did not look for Class B space in SF and were not facing an increase in rent with their expiring lease. It was a business decision for them to leave the City. The point being that, as you say, Class B tenants have many options and have a host of different business related factors that determine what they do when the lease is up. At least for now, it does not look as if Class B landlords with expiring leases have the leverage to ask for a rent increase.
It would be nice to know why there, seemingly, is not a push to convert existing Class B space to Class A (in certain parts of the city near the CBD) as there may be an excess of Class B space going forward.
“Given the significant drop in rents for Class B space”. What significant drop? What were the rents 2-3 years ago? I suspect that class B space is still doing great rentwize relative to a few years ago. Do you have stats?
As for B to A conversion, I suspect the cost and hassle is prohibitive on most buildings. Interior remodels aka lipstick on a pig is one thing (and many class B owners will do that during turnover.) But large open atriums, lots of glass, impressive entryway, etc. are usually not easily achieved in older buildings. Especially those drab and depressing 60-70’s office buildings.
On impact of tax bill, I agree w/ a few posters that for most buying $2m homes, it will have a negligible impact as reduced fed taxes on incomes in the higher bracket offset the impact of deductions and limits on SALT/Property Taxes. Also – it appears we are rapidly pushing thru legislation to circumvent this unjust tax plan on blue states by making state taxes charitable contributions (probably many years of legal war on that one but experts say it has and should stand up).
I’m looking to buy and I’m MUCH more worried about the increase in rates. They skyrocketed just this past week. Another 1/2 pt on a 7 figure loan is nothing to sneeze at. And I too am looking at all of this sq ft leased by blue chip tech companies and those going IPO wondering where will these workers live?
Regardless of they are w/ IPO companies, most have stock options and FB, Amazon, Salesforce are tracking on record highs (some of which are + 40% in the last year alone) and now held up higher by the corp tax cuts/repatriation.
Seems as soon as we start to see some softness – bc these $/sq ft comps are absolutely nuts – we see the trend line continue and creep higher. Very interesting to see what kind of inventory comes on in March when season kicks off.
Excellent reasoning wcsf,
I think raising rates represent much more downward pressure than changing tax rates. Maybe different market segments will be impacted differently.
– People buying for +$2M will hardly be deterred by a .25bps increase in interest rates. They probably own substantial assets which have benefited disproportionately from the stock market bonanza in 2017.
– People buying for <1.3M in the Sunsets, Bernals and Portolas of this town might be much more affected by interest rates. Most of them are young, working parents with solid jobs but limited assets and tons of variable expenses (nannys, preschools etc.). A 25bps increase on a $750'000 mortgage represents $1'800 higher payments per year. That can be enough to make people look at opportunities out of state.
Wake me up when the 12-mo. moving average is down more than 2%
Of course, while the market in San Francisco started turning down in the first half of 2006, the median sale price didn’t peak until 2007 and the median’s moving average wasn’t down over 2 percent until 2008 (at which point the weakest segments of the market had been on the decline for over two years).
Today, some are arguing the low-end segment is out-performing other segments. This would skew the median lower than that which would represent total market conditions (think “mix”). So maybe look for about a 5% decline. Or ignore that as the evidence of low-end out-performance is exceedingly weak.
Really, just look at Case Shiller, and you’ll have a very strong indicator of the market. Next, look at broad SF RE numbers — not just medians, but avg/sf, inventory numbers and trends, days-on-market, and many others which are widely available. Lastly, look at anecdotes, which are the weakest market indicator and easily manipulated. For example, realtors will generally only show you the strongest comps and ignore the weakest in trying to get you to pay more, and gets the sale made quicker which ups commissions.
Your mistake here is that you completely misunderstand the situation. The median is primarily a measure of what is currently selling *regardless* of it’s previous sale price/financial performance. The low end can outperform on an apples-to-apples basis and the median can still go up as long as there are more above median sales, even if those sales are at steep losses.
Many examples on this site show financial losses that raise the median. And DOM, price/list as reported by the industry are just garbage. You can see many many examples here of poor performing properties that nevertheless recorded as “over asking” and low DOM.
And you should take a look in the mirror regarding comps which are not “apples-to-apples” as it seems the editor regularly busts the sales you post on here feigning to be apples
Many properties are posted here *before* the outcome is known, defeating the claim of cherry picking. And even those which are posted retroactively seem to be thoroughly vetted with respect to changes in the property during the holding period. Which is a key advantage of looking at individual data points.
Bulk averages, such as Case-Shiller, generally do not have actual humans digging into each data point to check the property for non-market change (i.e. renovation, changes in zoning/permitting, TIC-to-condo,…). Case-Shiller uses heuristics based on unusual price changes and resale intervals to de-weight/exclude data points, but I don’t think they even use building permit data to flag transactions as being non apples-to-apples.
Though some other repeat sales indices do use building permit data, particularly in SF which sees a great deal of non-permitted work being performed, this data may not be all that reliable. And since in general people are performing renovations and/or land use changes to improve the value of their property, these changes are more likely to be in the upward direction than in the downward direction.
And lastly, there is a big difference between an anecdote, which is just an unverified story someone tells, and a data point, which is just one point of data, but nevertheless has been vetted for accuracy. A collection of anecdotes is a story book, a collection of data points is a trend.
Even with apples, or near apples, there are discrepancies that are difficult rationalize. For example 1071 Alabama just sold for $2.85. It’s only about 1800 sq ft, almost $1600 PSF. 1051 Alabama a few doors down sold for $2.4 in December 2017 and is larger at 2200 sq ft. Both homes renovated similarly/nicely imo. And compare that to nearby 948 Hampshire which sold for only $900+ PSF. All these homes renovated to similar standards. Crazy disparity, even if you account for the difference in home sizes.
Don’t confuse “apples-to-apples” with subjective “comps.” And keep in mind that 948 Hampshire, which just just sold for 16 percent less on an apples-to-apples basis, measures a rather non-comparable 3,500 square feet.
But the sale of newly renovated homes does help to push the median sale price up.
You are relying on the apple from the editor for this where he tend to choose the down vs. the apples to suit his narrative. Additionally re. Case Shiller what we don’t know is how they weight and what they consider unusual price jumps. Are they discounting large up (and down) apples based on a national average. Do they see a $300k gain loss as out of line and therefore not counted or discounted when here we know that is very possible? And then there are the properties like 1071 Alabama where it is a huge price but won’t be reflected due to the extensive remodel. The premier property in SF to many is the freshly remodel home, the value of which isn’t take into account by CaseShiller.
As much as some would like to rationalize our selection of apples as being nefariously “cherry picked” or driven by a desire to fit a predetermined narrative, they’re not.
Except for the rare occasions where there’s an outcome that’s particularly newsworthy, our apples are picked prior to knowing their outcomes and we actively avoid transactions which tend yield outlier results (such as inside sales or properties with fundamental changes which aren’t readily apparent).
And in fact, our only “narrative” remains a buy-side understanding of the market based on actual analytics and early trends, sans the typical industry spin.
Case-Shiller looks at the difference in price change for each particular property vs the local market average for that time period and de-weights the data point based on the absolute value of that difference (up or down). They state that typically 85-90% of transactions have weight 1, 5-8% between 1 and 0.5 and 5-8% between 0.5 and 0.
In addition to the weight assigned based on relative price change, another weight is assigned due to the time between repeat sales (holding period). Among other errors, this is meant to account for remodels/renovations which have not otherwise been detected. Any repeat sale within less than six months is completely excluded. Six month holds have weight one. With 10 year holds typically dropping to between .8 and .55.
Since they actually won a Nobel Prize, you can be sure they’ve dotted their ‘i’s and crossed their ‘t’s with respect to the math. It’s just that there is an inherent conflict between designing a system to collect and analyze bulk data for a whole nation vs dealing with all the idiosyncratic properties and neighborhoods of a town such as SF. Even simply lumping SF proper with the SF MSA has its issues. Of course as you point out, while human analysis can tease out more of the subtleties of the data, people are subject to their own biases and subjectivity.