5 ms·

You should look into how they calculate the numbers haha, it’s comical. They send letters to homeowners and ask them “How much do you think you could rent your

You should look into how they calculate the numbers haha, it’s comical. They send letters to homeowners and ask them “How much do you think you could rent your house for?” There’s no need for this idiocy I’m sure one of you could scrape the Zillow website in an afternoon and come up with something better.

LeifCarrotson 5y ago
It would be very easy to scrape Zillow in 2021, sure. But it would have been very difficult to do so in 1970!
- wonderwonder 5y ago
  sure but they could easily scrape Zillow in 2020 and compare the numbers to 2021. Those numbers would show a ~17% increase. If they are unable to adapt their model to modern data sets then the model is pretty much useless. Edit: General response to the responses. I am not advocating for Zillow to be a source of truth, I was responding to the parent comment of my post and their mention of Zillow. What I am saying is that we are not limited to using methodologies of the past, there are vast datasets and modern data collection methods available to us now that can be used.
  - Rebelgecko 5y ago
    Considering the financial issues Zillow is experiencing, I'm not sure that it makes sense to trust their pricing models. Zillow and Opendoor bouncing off of each other has given some neighborhoods the real estate equivalent of Amazon's book about flies that was algorithmically priced at $24 million (https://www.wired.com/2011/04/amazon-flies-24-million/ https://www.wired.com/2011/04/amazon-flies-24-million/)
    - Supermancho 5y ago
      > I'm not sure that it makes sense to trust their pricing models. Zillow, Redfin, etc list at prices that may be inflated, but they always list the purchased pricing as far back as they can and raise estimate to last sale if higher.
    - missedthecue 5y ago
      I don't think his point is Zillow specifically, but the housing market as it exists today. Zillow is his example.
    - SilasX 5y ago
      Zillow's model is fine in the aggregate, which is what the parent was suggesting it for. They ran into trouble profiting from it because adverse selection hits them hard -- they'll only be able to buy from the very homeowners whom the model most overpredicts the value of, because of unobservables it can't model. None of that relates the question of whether Zillow, on overage, correctly predicts how much the market, in the aggregate, has moved. I think you're confusing separate issues because they both merge into a narrative of "zillow bad".
  - joshuamorton 5y ago
    My rent dropped 20% (for the same unit) in the past year. That wouldn't be reflected on Zillow, but should be considered as part of the cpi, right?
    - kgc 5y ago
      Which city?
  - bosie 5y ago
    And when Zillow goes bust or fakes the data etc, what do you do?
seoaeu 5y ago
You are confusing "rent" and "imputed rent". The former is easy to calculate, and the BLM does calculate it accurately by simply asking a whole bunch of renters how much they pay each month. But imputed rent is a more subtle and difficult number. The goal is to know what the market rent would be for all the homes not currently on the rental market. I'm sure you can find other sources on why this value should be included in the CPI calculation (in addition to rent), but once you recognize that, it should be clearer why the Zillow approach wouldn't work: Zillow and similar sites only know about what homes sell for, or rent for. Now could an expert come up with a model of what a home might rent for given sale price, number of sqft, bedrooms, etc.? Of course. But doing so very much isn't an afternoon project because the data is tricky: the owner occupied homes differ from rental units in important and hard to evaluate measure. For instance, apartments around my city are more poorly maintained than similarly aged ownership units.
- missedthecue 5y ago
  >The goal is to know what the market rent would be for all the homes not currently on the rental market I don't understand why this is necessary, or why asking someone who hasn't been on the housing market for 15 years what the value of their home is might be good data collection.
  - seoaeu 5y ago
    I'll let someone else comment on the data collection part. But for why this is necessary, how else would you compute the housing costs of someone who purchased a home 15 years ago and hasn't moved since? That number is needed to compute the overall CPI, so you can't just skip it
    - missedthecue 5y ago
      Why can't you just skip it and calculate a median of all the nearby homes and rentals that have transacted on the market over the past x months? That seems more objective than asking people who have no idea what rent is what they'd rent their home for. I bet if you asked a panel of 1000 homeowners that haven't bought or rented in 10+ years what rent was going for in their cities, they'd undershoot it by 20% or more.

LeifCarrotson 5y ago

It would be very easy to scrape Zillow in 2021, sure. But it would have been very difficult to do so in 1970!

wonderwonder 5y ago
sure but they could easily scrape Zillow in 2020 and compare the numbers to 2021. Those numbers would show a ~17% increase. If they are unable to adapt their model to modern data sets then the model is pretty much useless. Edit: General response to the responses. I am not advocating for Zillow to be a source of truth, I was responding to the parent comment of my post and their mention of Zillow. What I am saying is that we are not limited to using methodologies of the past, there are vast datasets and modern data collection methods available to us now that can be used.
- Rebelgecko 5y ago
  Considering the financial issues Zillow is experiencing, I'm not sure that it makes sense to trust their pricing models. Zillow and Opendoor bouncing off of each other has given some neighborhoods the real estate equivalent of Amazon's book about flies that was algorithmically priced at $24 million (https://www.wired.com/2011/04/amazon-flies-24-million/ https://www.wired.com/2011/04/amazon-flies-24-million/)
  - Supermancho 5y ago
    > I'm not sure that it makes sense to trust their pricing models. Zillow, Redfin, etc list at prices that may be inflated, but they always list the purchased pricing as far back as they can and raise estimate to last sale if higher.
  - missedthecue 5y ago
    I don't think his point is Zillow specifically, but the housing market as it exists today. Zillow is his example.
  - SilasX 5y ago
    Zillow's model is fine in the aggregate, which is what the parent was suggesting it for. They ran into trouble profiting from it because adverse selection hits them hard -- they'll only be able to buy from the very homeowners whom the model most overpredicts the value of, because of unobservables it can't model. None of that relates the question of whether Zillow, on overage, correctly predicts how much the market, in the aggregate, has moved. I think you're confusing separate issues because they both merge into a narrative of "zillow bad".
- joshuamorton 5y ago
  My rent dropped 20% (for the same unit) in the past year. That wouldn't be reflected on Zillow, but should be considered as part of the cpi, right?
  - kgc 5y ago
    Which city?
- bosie 5y ago
  And when Zillow goes bust or fakes the data etc, what do you do?

seoaeu 5y ago

You are confusing "rent" and "imputed rent". The former is easy to calculate, and the BLM does calculate it accurately by simply asking a whole bunch of renters how much they pay each month. But imputed rent is a more subtle and difficult number. The goal is to know what the market rent would be for all the homes not currently on the rental market. I'm sure you can find other sources on why this value should be included in the CPI calculation (in addition to rent), but once you recognize that, it should be clearer why the Zillow approach wouldn't work: Zillow and similar sites only know about what homes sell for, or rent for. Now could an expert come up with a model of what a home might rent for given sale price, number of sqft, bedrooms, etc.? Of course. But doing so very much isn't an afternoon project because the data is tricky: the owner occupied homes differ from rental units in important and hard to evaluate measure. For instance, apartments around my city are more poorly maintained than similarly aged ownership units.

missedthecue 5y ago
>The goal is to know what the market rent would be for all the homes not currently on the rental market I don't understand why this is necessary, or why asking someone who hasn't been on the housing market for 15 years what the value of their home is might be good data collection.
- seoaeu 5y ago
  I'll let someone else comment on the data collection part. But for why this is necessary, how else would you compute the housing costs of someone who purchased a home 15 years ago and hasn't moved since? That number is needed to compute the overall CPI, so you can't just skip it
  - missedthecue 5y ago
    Why can't you just skip it and calculate a median of all the nearby homes and rentals that have transacted on the market over the past x months? That seems more objective than asking people who have no idea what rent is what they'd rent their home for. I bet if you asked a panel of 1000 homeowners that haven't bought or rented in 10+ years what rent was going for in their cities, they'd undershoot it by 20% or more.