While there are roughly 60 percent more homes in contract across San Francisco (500) than there were at the same time last year, there are still 50 percent mores on the market (940) and 130 percent more listings for sub-million dollar homes (350).

And while the “expectation gap” between the average asking price per square foot of the homes which are still on the market ($1,027) and the average price per square foot of the homes which are in contract ($972) has ticked down to around 5 percent, the average list price of the homes which are in escrow is down around 4 percent versus the same time last year and there are twice as many reduced listings (180) on the MLS.

29 thoughts on “Contract Activity Up Along with Inventory, Expectation Gap Down”
    1. No, all assets are going to continue to go up in value as the dollar gets crushed. Joe and Jerome just printed 1.9TT and this hurts the people they are trying to help. Dollar devaluation only helps those who have assets/real estate.

      1. That 2T is going mainly to people earning less than 75k and only ~1400 bucks plus some for the kids. That wont allow people into the million dollar condo market here. Don’t see that having an affect. What will have an affect is the 120billion+ a month of bond bailouts and 0% rates. Implied inflation rate should be around 4% based on commodity boom so your bank account is earning -4% interest.

        1. The 2TT is already spent (on rent, iphones, food and shares on the Robinhood app) That helps people who own Real Estate, AAPL, Kroger and other Equities.

          The Feds monetary policy exacerbates this excess. DOGE going up is a symptom of all this (FYI, I am a BTC bull)

        2. Unfortunately, it will be a K shaped recovery. Those with assets and those without. Those who benefited greatly from the stock market will have plenty of money to spent on real estate, in fact, a lot of them are thinking about take some of the gains to put into real estate to have a more balanced portfolio.

      2. I’m not so sure. I agree that we are probably looking at a great deal of stimulus coming down the pipe and that is going to prop up asset prices in the aggregate. But I’m not sure every single asset is going to benefit equally. We already are luxury housing markets booming across the country in general, while SF, NYC and other urban centers are seeing weakness. For RE that you plan to occupy, cheap money lets you bid up RE in general but it doesn’t tell you where you want that RE to be. Monetary policy is a blunt tool.

        For investments, I recall that ‘ST’ said that when he bought distressed property in prior cycles even at distressed prices he was buying at negative cash flow and hoping for appreciation. Some other posted compared that to other parts of the country (Kansas CIty I think) where you were buying cash flow right away.

        Buying into a loss with cheap money is still buying into a loss. I don’t think the ‘buy at a loss and hope for things to get better’ assets are going to go in the same direction as the ‘buy into cash flow’ assets.

        When I look at things like that last nob hill apple that went for 17% below 2015 pricing despite all the stimulus since then. 9% inflation and much lower (25%? by my eyeballing the graph) interest rates. I think that the volatility of some RE is going to greatly outweigh any stimulus.

        1. There’s a difference between “buying into a loss” with a property that’s performing more or less as well as it’s going to perform, and buying a distressed property that can perform much better after some TLC (often expensive TLC) and better management. I’d argue that buying a property which is likely to keep losing money, in hopes of making money on appreciation, is more like gambling than investing.

          Buying a property which is underperforming, and then solving its problems and getting it to perform, that’s “value add” investing and is something I’ve done well with. But I’m not sure which scenario ‘ST’ was referring to.

          I buy in parts of the country where, if a property isn’t cash flowing at purchase, it’s because it’s got problems you need to solve. You make your money by improving the property’s performance, and if you get some appreciation too, that’s a nice bonus. It’s hard work and doesn’t happen overnight.

          Most of the country is not like SF, where people expect to buy a property and have it double in value in a few years with little effort required. I’ve done well with property here too, but those successes feel more like good luck/timing than a reward for brilliance or effort expended.

          This is a long-winded attempt to agree with you, that “buying into a loss” with no clear path to cashflow is quite the gamble these days, in any days, really. When that volatility comes it may crush the dreams of some of the speculators whose hopes (and money) were pinned on appreciation that may be a long time coming.

          1. I agree that when you are talking about actively making improvements that’s a different story. But with the posters talking about Fed and/or fiscal policy pushing up asset prices that is inherently a passive strategy.

      3. Well said. All this stimulus is creating in-equality. Those who already have RE or stock stand to benefit the most, where the real needy get $1400 that lasts 2 weeks…

  1. You already missed that bottom around Nov time frame last year. With unlimited amount of money printing and free cash for all, asset prices – across all class – will continue to go higher for the next 1-2 years.

    Beyond 1-2 years, though, it is hard to imagine what is going to happen. Stagnation?

    1. Based only on my own anecdotal browsing, I think the bottom was actually closer to May 2020. By November, prices had recovered and were up YoY.

      Now is there another dip coming in the next 1-2 years? Maybe, maybe not.

      1. In addition to being down around 4 percent on a year-over-year basis, the average price per square foot of the homes which are in contract in San Francisco is currently 6 percent lower than in May. And on an apples-to-apples basis, we haven’t seen a recovery.

        1. Looking only at YoY data means you will only know when we hit a local minimum one year after the fact. Plus that we are looking at monthly numbers and transactions take one month to close, meaning you are potentially 14 months removed from the true moment of the minimum by the time you would detect it. Average price per square foot is subject to mix shift (it reflects what’s selling not how prices evolve) and should never be used as a price index. Apples to apples is the true way to go but you have to use a case Schiller approach for SF-only data to measure QoQ trends.

          1. “Apples to apples is the true way to go but you have to use a case Schiller approach for SF-only data to measure QoQ trends.”

            That would be truly ideal, but is there even enough source data for that? Not that many transactions in SF city and not all of them are apples. And you’d need those with a range of sale dates to make a quarterly index.

            You’d probably want to separate out conods vs SFHs since condos look like they led the way down… I’d say that anecdotally SFHs in the 4+ Br range are probably doing better then most, but they seem pretty rare.

            It would be great to have a statistically valid price index for all the relevant housing types and neighborhoods in SF, but I’m just not sure that’s possible. I think the best you can do is look at the apples as they come in and make a judgement call.

          2. I actually did just that for NYC (see name link), the data there is readily available. Q4 showed 1% QoQ gain from Q3. I bet someone could do the same for SF if you just found the data on property transactions (which I believe the editor has – maybe they can share the source and I can put it together)

          3. We sincerely appreciate the approach and applaud the output and additional market insight.

            But as noted, your index doesn’t differentiate between movements in the market and changes in value due to properties being “remodeled or otherwise improved” between sales, which can be particularly acute in neighborhoods that are gentrifying and/or have a higher percentage of older housing stock and fails to yield true “apples-to-apples” sale pairs and a resultant index.

          4. Case Shiller has been showing top tier gains for the region for some time now. Case Shiller is comprised of apples.

            But again, the flaw with apples is that properties age, get outmoded, have problems develop over time. Case Shiller does not account for these things.

            They’re all data points. Nothing is exclusively definitive, apples included. Equally, the outright dismissal of remodeled properties as data points is also specious. Why? Take them in aggregate for an area together and consider that data point.

          5. Please don’t mix condos and SFRs as a rebuttal. Condo demand in a pandemic truly is “different this time.” Thanks in advance.

          6. @asdf – That is very nice. I would like to see something like that for SF as well. The editor has a very good point regarding errors (remodeling, non arms length, …) in the raw data. But putting my earlier question quantitatively about the statistical accuracy of the algorithm, What size of error bars do you get for a city size data set? Looking through your link it looks like you have a number of neighborhoods with 10’s-20’s of data points. I’d assume that the Case-Shiller people have some way to generate error estimates, but I personally don’t know enough about the algorithm to even have a ballpark idea of how data set size translates into error range.

            Just thinking about it a bit it seems hard to ballpark. If you have 10 transactions that were bought in Q3 2020 and sold in Q4 2020, that gives you a lot of information about the Q3-Q4 change. But if you instead have 10 Q4 2020 sales that were all bought in Q4 2015, that would seem to give you little or no information about 2020 Q3-Q4 change. I’d think that not just the number of data points, but the distribution of buy/sell dates matters for generating an error estimate,

            But whatever the error is, I’d assume that more data leads to less error. So that would lead people to aggregate neighborhoods/price ranges/ property types to increase the data set size. But based on your results (and many anecdotes) there is some pretty large divergence of the price trajectory for different neighborhoods, property types, price levels,… If I’m reading your results right you show a -27% YoY for the financial district and a +18% YoY for East Village/Alphabet city.

            Even for QoQ it looks like you show a range of -15% for Financial to +6% for Crown Heights.

          7. @Ohlone Californio – “Case Shiller has been showing top tier gains for the region for some time now. Case Shiller is comprised of apples.”

            The main problem with regional Case Shiller is that it is aggregates many RE markets/sub-markets that are going in different directions. Look at the data Asdf ran for NYC, It looks like there is a 45% spread between the best and worst performing neighborhoods. And that is just neighborhoods within NYC. With regional CS you are aggregating whole cities with widely different characteristics.

            It’s true that properties age, but that isn’t the problem you think it is. Properties aged 2 years from 2004-2006 and the same two years passed between 2008-2010. But in one case the market also went up over those two years while in the other it went down. And look at neighborhoods too. One year passed in one area and prices went up 18% and the same year passed in another and prices dropped -27%.

          8. @Ohlone Californio – No one is outright dismissing remodeled properties. But when looking at remodeled re-sales that is even more a case that you just need to look at them and make a judgement call about the value of the remodel vs change in the market. If you wanted to use remodeled properties in a more statistical way, you’d need to create some model to factor out the value add vs market change. This would seem hard to do with great accuracy, especially in markets such as SF and NYC with very bespoke properties. Even gathering the input data of what remodeling work was actually done seems challenging when it seems that a lot of work in SF is un/under-permitted.

          9. The problem I think it is is not the problem you think it is, especially considering longer timeframes than the 2 year ones you’ve posited. Styles, tastes, wear and tear, functionality, these things can really change in five, six years even.

            Like I said you can pick apart anything. But Case Shiller top tier — largely conversant with SF prices — has been up month in and month out. So has $/ft year over year comparisons. Neither are perfect of course. What I don’t see on here are bearish posters analyzing what’s wrong with taking singular hand selected apple(s) and extrapolate that outwardly, broadly painting.

            I don’t agree with the unpermitted and under-permitted work take at all. That’s a really stupid game to play in this day and age of easy whistle blowing, particularly at the high end.

          10. But 2 years vs 5 years doesn’t matter. 5 years can pass during good times and 5 years can pass during bad times. 5 years can pass in hot neighborhood 5 years can pass in neighborhoods gone cold. Time passes everywhere and all the time. Yet in some times and places apples (and indices based on apples) go up and and some times and places they go down. Pointing out the passage of time doesn’t add any real insight to whats happening with the RE market.

          11. That’s not what I said though, merely pointing out the passing of time? You bring up an interesting point when it comes to neighborhoods. Though I think when it comes to SF and general trends in neighborhoods + longer time frames? I think you’d struggle to find evidence of a neighborhood gone cold. At least over the past four cycles or so.

          12. “I think you’d struggle to find evidence of a neighborhood gone cold”

            The frozen listings on zillow and c-list say that SOMA’s a Sub-Zero right now.

          13. Indeed. That’s something, SOMA condos performance thus far during the pandemic. Not what I was getting at but fair enough.

      2. Thanks you for the stats (seriously)! Is the trend the same in the east bay? Oakland and Berkeley are where I have been looking.

  2. Also interestingly, in the info-graphic on the SF tech exodus I posted on the other thread they referenced USPS mail forwarding data. I initially though this was the same analysis that was dinged here a few months back for not providing YoY or inbound data for context. But it appears that the same source has updated their analysis to account for both YoY comparisons between 2019 and 2020 and for inbound migration.

    The TLDR is that the number of outgoing USPS forwarding requests for 2020 was about 5x the amount for 2019!

    The also get some commentary about this from SF’s chief economist:

    “Based on what the postal service says of the total numbers in 2019 and 2020, it looks like there were about 100,000 more requests over those 8 months, which is a big number,” said Ted Egan, San Francisco’s chief economist. “That shows a clear increase in outmigration, and that lines up with other data we’re seeing.”

    And provide some insight as to why some of the official population counts based on DMV data might be lagging:

    “But given the effects of the pandemic, those population estimates— which are based on Department of Motor Vehicles (DMV) registrations, Internal Revenue Service (IRS) returns and other indicators— are likely to change considerably as new information comes in, said Phuong Nguyen, a researcher at the California Department of Finance.”

    “The data we collected is very preliminary. Because of COVID, some places might not have up-to-date data, or a slowdown in data processing.” Nguyen said. She noted that the DMV’s closure between March and May 2020, along with extensions of tax deadlines, could spell delays in accurate population counts at the county level.”

    1. I would think DMV data would be among the least useful, since we’re talking – obsessing?? – about young, single, apartment/condo/dormitory dwelling people, a group that might aptly be described as “carless”.

      That having been said, even more refined – or less “flawed” in the vernacular preferred by some – procedures can go awry: the Department of Finance’s estimates seem to have a consistent bias toward overstating populations (at least against the Census figures that come out afterward).

Leave a Reply

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