Thursday, March 19, 2009

EstiMike Blog

Bubble Meter has new competition.

14 comments:

  1. There is a ton of real estate info on the web, but it's hard to find unbiased information. Most info comes from someone with a vested interest like a real estate agent. Even sites like Zillow are tightly linked to real estate agents. Thanks for providing useful, unbiased info, and I hope I'll do the same on my EstiMike site. Cheers, Mike.

    ReplyDelete
  2. That estimator leaves a lot to be desired. It's first major flaw is it's assumption that the price previously paid was a "correct" price reflecting the underlying value of the property. Second, it does not account for quality changes since the last purchase. Third, it does not account for quality changes in the neighborhood since the last purchase. Fourth, it's valuation methodology is based on the price index rather than on an income or rental index. Etc, Etc.

    ReplyDelete
  3. 1) Things like feel, quality and value are a matter of opinion.

    2) As far as its valuation method being based on price, so what? Rentals and incomes havent shot up 200%-400% in the past 8 years, prices of homes have!

    ReplyDelete
  4. John Fontain said...
    That estimator leaves a lot to be desired. It's first major flaw is it's assumption that the price previously paid was a "correct" price reflecting the underlying value of the property.

    EstiMike goes out of its way to explain its own shortcomings. It's just one tool of many to estimate a home's value. It does not assume the previous price reflected the underlying value of the property. Instead, it assumes the previous price was the proper market price, which is usually a reasonable assumption.

    Second, it does not account for quality changes since the last purchase. Third, it does not account for quality changes in the neighborhood since the last purchase.

    Most of the time houses and neighborhoods have very few significant quality changes within a ten year period. Homeowners tend to overestimate the marketable value of their home improvements.

    Fourth, it's valuation methodology is based on the price index rather than on an income or rental index. Etc, Etc.

    A rental or income index would be the incorrect tool to use. If someone bought a house at the bubble peak and then tried estimating its current value based on the change in a rental index, he would be trying to sell at a price ABOVE the bubble peak price. That would be downright foolish. A price index is a far better tool to use when trying to determine the MARKET price of a house.

    A buyer trying to avoid buying during a bubble may want to look at a house's cap rate.

    ReplyDelete
  5. james, how could using a rental index at the peak result in a price above the bubble price? given that rent prices did not increase as much as the purchase price, valuing the property based on rental cash flow would result in a price far below the asking price at the time.

    ReplyDelete
  6. james, the problem with the estimike is that, in assuming "the previous price was the proper market price," it:

    a. assumes the previous price was the proper market price

    and more importantly,

    b. assumes the previous price reflected the house's value.

    If there is one thing that become glaringly obvious during the housing bubble, it was the fact that prices diverged wildly from underlying values.

    My point is that a calculator that attempts to tell you what a house is really worth, and yet does so without regard to any measure of worth but instead only uses what we all know were wildly incorrect market prices, is inherently flawed.

    IMO, the estimke is completely worthless. Much better to use historical price to rent ratios and estimated rents of your target property to guage its worth. Can be done with simple math and in only a few minutes.

    ReplyDelete
  7. Oh, and one more big flaw is that it doesn't account for variances in price changes within a certain region. For example, let's say case shiller for DC is down 25% from 2005. So you put in the price of a property that was purchased in 2005 for $400,000. The calculator will tell you that it is worth $300,000. But if the house was in Prince William County, current market prices would say the house was "worth" only $200,000.

    ReplyDelete
  8. mrowens said...
    "james, how could using a rental index at the peak result in a price above the bubble price? given that rent prices did not increase as much as the purchase price, valuing the property based on rental cash flow would result in a price far below the asking price at the time."

    The EstiMike site works by taking the last purchase price and adjusting it by the change in the S&P/Case-Shiller Home Price Index for the local metro area, to arrive at the current MARKET price. Like home appraisers and Zillow.com, EstiMike is NOT trying to determine the intrinsic value of the home, just the market price.

    Let's use a simple example of how it works, based on the national median home price. Let's say that in August of 2005, a home sold for $235,000. Nationally, home prices have fallen 24% since then, but rents have INCREASED about 14%. If you adjust the $235,000 by the change in home prices, as EstiMike does, the house would be valued at $178,600 today. That's a $56,400 decline. However, if you follow John Fontain's suggestion and adjust by the change in rents since then, the house would be valued at $267,900 today. That's an INCREASE of $32,900 since the bubble peak! Only a fool would pay 14% MORE when houses everywhere are FALLING in value.

    ReplyDelete
  9. John Fontain said...
    "Oh, and one more big flaw is that it doesn't account for variances in price changes within a certain region."

    You are correct here. I thought he said this on his website, but I don't see it.

    John Fontain said...
    james, the problem with the estimike is that ... it:

    a. assumes the previous price was the proper market price

    and more importantly,

    b. assumes the previous price reflected the house's value.


    EstiMike is very open about the fact that its valuation method has shortcomings.

    Regarding point (a), supply and demand ensures that most houses sell within a reasonable range of their market value. Price a home too low and buyers engage in bidding wars. Price a home too high and nobody makes an offer. Thus, using the previous sale price is a good starting point much of the time, but certainly not all the time.

    Regarding (b), I understand your point, but I don't believe EstiMike is making any claims regarding underlying value. Like any method that judges price based on comps, he is trying to estimate the MARKET PRICE, not the intrinsic value.

    People like you and me, who believe that we should buy based on underlying value, simply aren't buying now. I think what you really are suggesting is what I often do: Figure out the pre-bubble (late 1990s) market price, then adjust by the change in either rent or overall inflation since then.

    ReplyDelete
  10. I second everything that James said. John's point about local variation is well taken; if you have a particular town that is out-of-step with the overall metro region, then EstiMike may not give accurate results. I'd love to see some specific data, but I personally believe it's not too common. Furthermore, EstiMike uses the tiered Case-Shiller data (when available), which should reduce this issue to some degree.

    Getting a bit technical, the Case-Shiller data is available in more narrow geographies, but this data is unpublished. I have considered getting the data from Fiserv, but using private data would defeat some of the purpose of the EstiMike site.

    Again, using Case-Shiller data is a tool to motivate discussion and price negotiation. You still need to use common sense to understand the specific condition of the property and the neighborhood.

    ReplyDelete
  11. "Mike Said...

    if you have a particular town that is out-of-step with the overall metro region, then EstiMike may not give accurate results. I'd love to see some specific data, but I personally believe it's not too common."

    Actually, case-shiller themselves have come out and said this phenomenon has been seen in several cities. DC, SF, Boston, LA, etc. All have huge declines in exurban areas and minimal losses in core areas:

    http://www2.standardandpoors.com/spf/pdf/index/052708_Housing_bubbles_collapse.pdf

    Median data says core DC is down 5% from the peak, while exurban prince william is down 40% or more

    The CS tiered index doesnt really account for this since there is plenty of high end stuff in the Exurbs that has crashed, and even low end stuff in the city that has not.

    If you could get the locally based data from Case Shiller you would get a much much more accurate estimator.

    ReplyDelete
  12. james said: "However, if you follow John Fontain's suggestion and adjust by the change in rents since then, the house would be valued at $267,900 today. That's an INCREASE of $32,900 since the bubble peak! Only a fool would pay 14% MORE when houses everywhere are FALLING in value."

    I may not have been clear enough in my original suggestion. I don't mean to apply rent changes since the peak to peak prices to determine current market "price." I meant, but did not write clearly or in sufficient detail, that one could determine "value" based on pre-bubble prices adjusted for rent inflation.

    I'm trying to provide a correct model for valuing a housing, not pricing one. To price a house, one need not use estimike, one need only look to recent comparable sales (something that would be much more accurate in estimating market price for a specific property than applying an entire market C/S index to a previous purchase price of a specific property).

    And again, computing market price is mostly a worthless effort no matter how it is done. People should be focused on computing value and only buying when they get more in value than they pay in price.

    If estimike wanted to create a tool that would really be useful and help people, he should consider a value calculator based on historical cap rates (also known as price to rent ratios). He could take pre-bubble historical price to rent data by city and have the user input the amount of rent that could be reasonably generated by a property and then compute its true worth. Now that would be something that would be truly helpful, because it would help those that use it to avoid drastically overpaying for a house.

    ReplyDelete
  13. Thanks for the suggestions.

    I agree that the more localized Case-Shiller data would improve the accuracy of the prediction, but it would also make EstiMike no longer transparent because it would use data that is not available to the general public.

    As for the second suggestion of computing value, this is interesting but currently out-of-scope for the EstiMike site. EstiMike determines market price. Whether you think that represents a fair value for the property is a different question, and I suspect that this question would generate a lot of controversy - especially for real estate agents who have a vested interest.

    Using comps to determine the market price is more art than science because it requires you to judge which properties are comparable. Even similar homes inside a neighborhood can have different values due to condition, upgrades, renovations, noise, views, power lines, neighbors and many other specific factors. The good part about using the Case-Shiller index to predict prices is that most of these factors (except condition and upgrades) are already considered in the previous sale price.

    There's one guy on the internet who is waging a personal war against Zillow for these various reasons. You might want to read what I wrote about this on the EstiMike blog.

    ReplyDelete
  14. It's good to know there's one person out there offering unbiased information on housing bubbles. I enjoy your polls, and have been reading the blog for a little while, and just wanted to comment.

    ReplyDelete