Are Internet Home Valuation Tools Worth Using? 12 Months of Historical Data

If you like to keep up with my net worth updates, you’ll know that I’ve been trying to estimate my home’s value using four different internet valuation tools:

It’s finally been a year since I started using these tools, and I wanted to look back on how successful this experiment has been. Some give one numerical estimate and some give ranges. When I got a range of values, I simply used the midpoint. Here are the results over the past year:

Accuracy. We all know the real value of a house is what someone is willing to pay for it. (Actually, it’s what a bank is willing to lend that person…) But as a group, I would say these services can’t possibly be accurate because they all seem to disagree all the time. In nearly every month, estimates would differ by $50,000 or more. Half are up over the past year, and the other half are down.

Even though they all seem to use similar housing information databases, all of them must interpret the data differently. Some may take into account the change in value of all houses in zip code, or the city in general. Or they may differ in what properties they consider to be comparable sales (by zip code, by distance, size, or age). How do you weigh different physical factors (i.e. number of bedrooms, bathrooms, and square footage)?

Volatility. Bank of America appears to have had the least volatility over the last year, while Coldwell Banker has the wildest swings. I believe the recent upswings are due to some more houses being sold in my area at the higher range recently, while for a time very few houses were being sold and those were at the low end. If so, it would follow that Coldwell Banker weighs recent sales more heavily.

Should I Keep Using These Tools?

Even though the house you live in might not be considered an investment per se, I still think one should track it’s value. The fact is that $200,000 in home equity here could buy me an entire house elsewhere if I were to move. (And I know people who have done just that.) I have seen that home price trends can vary across different neighborhoods within the same area, so you can’t just go by city-wide data.

However, due to the lack of reliable estimates from these internet valuation tools, I’m going to stop using them in my net worth updates. I initially tried to deal with the variations by taking the average of all four and also subtracting 5% to be conservative and 6% for real estate agent commissions, but the swings still dwarfed all my other accounts. I’m still undecided about what alternative to use.

Comments

  1. Can you get comps from a realtor or track sales through an MLS website periodically and estimate?

    Previously I was in a condo building and currently have a town house, so I’m spoiled as far as ease of estimating what my property is worth – I’ve always just used actual sales data for equivalent properties.

  2. @DB – I could pull comps myself, and I do check periodically, but I might have some personal bias as to what I consider actually “comparable” to my house. ;)

    In the end, maybe I could do that once every 6 months or so. I probably don’t care enough to do it monthly.

  3. More interesting to me is how all the services valued your house in about a $80,000 window one year ago and now their range varies by about $450,000.

  4. Why does your home value matter so much? In my net worth statements I track the things I can with dollar amounts. Other things are left as assumptions like home value, retirement tax bracket, social security, pension.

    These things are simply unknown until the need arises. Tracking that ghost value offers no real benefit to you until you really need it for a refinance or to sell the house. But then you will use much more accurate gauges like an appraisal rather than relying on Zillow or Coldwell Banker. So why track it at all?

    If you must track it take a good guess, update it every one or 2 years and be done with it. This situation reminds me of running retirement analysis on stock returns based on minutes of the day.

    I believe it was Jeremy Grantham who described the market as a man walking uphill using a yo-yo. He reminded us to watch the man, not the yo-yo. You’re watching the yo-yo with your house values :)
    Tom

  5. Even though real estate is technically an investment and part of one’s net worth, I think it resides (ha ha) in its own category.

    When I bought my house I researched the neighborhood, school system, etc, to make sure I was buying into an area where property values would be maintained. But the primary reason I bought my house was to live in it. It has a practical use, unlike a stock or bond, and therefore is not solely an investment vehicle.

    Speaking of vehicles, one could technically include cars/trucks/RVs as depreciating assets in total net worth. But most people buy these to use them and drive them, not as investments.

    It’s nice to see one’s entire net worth, including real estate, sometimes (like maybe once a year or every other year) but for monthly budgeting and cash flow purposes I would say it’s irrelevant, IMHO.

  6. Although it’s hard to swallow, I would use the most conservative estimate out of the bunch. Keep the surprise on the positive side. Incidentally when I had a refi in the spring, Cyberhomes’ estimate was the closest to the appraisal value.

  7. In my opinion Zillow seems to overestimate house prices. How about Trulia?

  8. I estimate my house value on a annual basis on the anniversary which I bought it. IMHO, rying to estimate on a monthly basis is too frequent.

    I basically looked at what other houses are selling for in my area and then conservatively estimate how much I think that I could get for my house minus the realtor commission. I have looked at some of the sites you mention, but I find that they are trying to apply an algorithm without any knowledge of my local market. For example I have done extensive renovations on my house, which is not accounted for on the websites that are estimating the house value.

    Since I reevaluate on an annual basis, I don’t get massive changes in my net worth. Realistically the house value only needs to be accurate if you plan to sell in the near future, otherwise a basic ballpark would suffice.

    My 2 cents…

    Tap

  9. seems that the volatility is the main concern: how about a 12 month moving average from the 4 sources? will definitely be less volatile, and should be conservative if housing prices are increasing over the long term.

    could also consider using the case-shiller index for your area from the initial purchase price but again it would not reflect neighborhood changes.

    last idea: keep the home price constant (or scaled for inflation) for a few years till the housing market is more stable

  10. I like Craig’s Idea – why do it everyone month? It is a home that you aren’t selling any time soon. Just throw it in at X and leave it X and updated it once or twice a year. While doing this, you can decrease your liability on the home from your mortgage statement.

  11. I personally am very concerned with keeping track of my home’s value. I have always tracked comparable sales. Immediate neighborhood – same houses. IT’s easy and I personally have found I have a better idea of FMV of my own house than most any realtor. I’ve had a few appraisals done in my life (refis) and I didn’t feel they were rocket science – were withing dollars of my estimates.

    I’ve said before, for Net Worth I go by assessed value, to show modest return on investment. (When my house was worth $600k+ I put $300 on my balance sheet. I didn’t think it was sustainable. Around that time is when I switched to the assessed method. Our assessment goes up 2% every year, that’s it by law). Thing is, our assessment will go down to market value if the value of our house drops. I do track that. It’s a way to not get over inflated asset numbers, WHILE keeping track if the house does lose value.

    I go with assessed value every year. I only update my net worth annually for house value. Though I admit it’s a bit of a hobby to watch neighborhood sales figures throughout the year.

    I haven’t been in a position to have my assessment lowered yet. Potentially in the next year. I am assuming I will agree with the assessment, for my net worth statement. Will see.

  12. I would just carry it at what you purchased it for plus any additions/improvments at FMV. I would only adjust it downwards unless you KNOW that it’s worth less.
    This is a bit accounting-ish, but there’s no real way to know unless you get official appraisals, and even those have been wonky lately. I don’t think adjusting the value monthly adds any real value to your financial statements and it just clouds the manageable change in your net worth. Once markets stabilize in a few years, I’d adjust it then, but until then, to hell with it.

  13. I think I mentioned this before — but I feel like the best way to value your house is to figure out what it would rent for (which is relatively easy compared to what it would sell for), divide by 6 (arbitrary number, but based on the percent that 30 year mortgage rates always circulate around), and multiply by 1000. For example, my house would rent for about $2400/per month, and therefore is worth approximately $400K, which is very close to accurate (based on nearby sales of similar houses). It has ranged in the last 5 years between $400K and $500K, but is now at the low end. I’d be curious to know what your number would come out to be.

  14. To keep things simple I use the OFHEO (now fhfa) housing index. It comes out quarterly for a couple hundred metro areas and yearly for another hundered or so. On the yearly anniversary of my house purchase I increase or decrease the value of the home (started with my purchase price 6 years ago) using the index (I actually divide it by 12 and retroactively apply it to each month of the past year so there are no spikes)

    I like this index for this because it is not tracking an average of all home prices (which is always skewed by how expensive or inexpensively people are building new homes) but is supposed to follow the price value change of an existing home.

    I don’t use my “net worth” number much and don’t think it tells you much over periods of time shorter than a few years, so the ease of plugging in this number once a year and the reasonableness of the numbers I’ve been getting compared to what I would “guess” my home is worth makes this a good enough process for me.

  15. I also like Craig’s idea of using a rolling avg price. I would suggest using a 6 month avg and then also remove the Caldwell Banker service since it is clearly much more volatile then the others.

    It seems slightly crazy how much the value of your house has moved over the last year, but in all likelihood it is a reasonably accurate portrayal considering how much all housing prices hve moved this last year.

    I would continue to track it, if simply for our enjoyment.

  16. BTW, the direction is what is most important, not the value. Cyberhomes is the most conservative, so I’d use that one, if you HAD to use these tools.

  17. I think the tools can be useful for a ballpark estimate. They are a quick and easy way to get a general range of the value of a property. To get a more accurate estimate you’d have to get a real appraisal. If your purpose is to just get a ballpark guess on the price then these tools are fine. If you more dependable value then you should get a real appraisal. Generally you don’t *need* a dependable value unless you’re planning on selling or refinancing in which case you’ll probably be required to get an appraisal anyway.

    Looking at the individual sites it seems that some have newer comps than others and that impacts the values. I think if the site uses a range of values thats much better since you have fairly decent expectation the real value is somewhere within that range.

    Zillow has a page with data on their accuracy:
    http://www.zillow.com/howto/Da.....curacy.htm

    Their estimates are off by 11% median nationally but it varies from city to city. About 2/3 of the time they’re off by 20%.

  18. I have to disagree with the people saying to just use cyberhomes, this may be anecdotal, but they seem to have some seriously bad methodology.

    I got a refi in February, and the appraisal listed my house value at $400K. I live in a neighborhood of near-identical condo townhouses, which are all currently selling for over $400K. My neighbor’s house is slightly nicer than mine and just sold (in April for $500K). Given the hundreds of near-identical houses in my neighborhood, my house value is pretty easy to estimate. The Zestimate for my house is $444K, which I think is high, but probably something like $410K is pretty reasonable.

    Cyberhomes however compared my house to numerous (less-nice) houses within a mile or two of my house, and compared it to exactly none of the hundreds of identical houses right next to mine. They did this and got the estimate $265K. Completely, utterly wrong and pretty strong evidence that their method of finding comparables stinks.

  19. How about doing what Robert Shiller does with P/E ratios, and taking 10 year averages instead of taking “today’s” value?

    The benefit will be that you won’t reflect wild swings in the valuation either going up or going down.

    Of course, it’s not as effective in these housing market, because your historic valuation will be still inflated due to bubble prices. In that case, one other way is to just take the price around 10 years ago, and do compounding at the rate of “inflation + 1%”. This will yield you pretty conservative price for the house. If people had done this, there wouldn’t have been any bubble in the US :-).

    (where do I get “inflation + 1%” from? : in Robert Shiller’s book, he presets his finding that the house prices in the US have appreciated by inflation + 0.2 % over the long periods (50+ years). The guy has been amazingly right about both tech and housing bubbles, so I trust him more than most economists.

  20. I think pulling good (sold!) comps is probably the best way to get an estimate of how much your house *might* sell for, but that too can be problematic, especially if you’re in an older neighborhood or one without a great deal of turnover.

    Personally, I’d probably just use the average price per square foot for sold homes in the area and use that number to calculate what your house is ‘worth,’ though again, that could be tricky if there aren’t many houses selling.

    I’d also just leave the number as is — don’t see a point in including the 5% ‘buffer’ (if you think your number is high, then there’s something wrong with your methods) or the 6% in commissions (are you including sales commissions or tax rates in any of your investment accounts, for instance?).

  21. Why not just use your property tax assessment value? That value will only change once per year, and it is probably limited in the amount it will change. This will force you to make infrequent updates to your house value, and usually (at least where I live) the tax values are usually conservative compared to what you would likely get if you sold the house.

    By doing this, it will make your net worth comparisons less tied to your house value, but will still provide for adjustments made only once per year in case of major changes.

  22. When Zillow says I’m $400,000 richer in the past 2-3 months (I have a post on this with two charts highlighting the history of my properties), you KNOW it’s hogwash! Don’t trust anything.

    Take all your illiquid assets, and THROW THEM OUT THE WINDOW when calculating your net worth. Too many people count on these assets and develop a false sense of security.

    If and when you retire, these assets are still there, GREAT! Just don’t count on it.

  23. all these tools are equivalent to trying to catch a mouse with a bear trap. I would just take the 3 that have most accurately reflected actual value and average them.

  24. MyCrossoverPoint says:

    I have had similar trouble tracking my home’s value for net worth purposes, and also had the problem that the volatile swings in its estimation dwarf the other increases in investments, etc.

    I had been using Zillow, but I’m thinking of adopting the approach some have mentioned of just leaving it alone and not trying to update it monthly.

    I like mimi’s idea, but that produces the problem of estimating rents, and I wonder if those estimates are as variable and volatile. mimi, I would be interested to know what resources you would use to estimate rent?

    What I have ended up doing is calculating my net worth by tracking four different categories: assets (house, car, etc.), cash & investments (or liquid assets), retirement, and liabilities. Ultimately, what I want to see is my cash & investments increasing and my liabilities decreasing. Retirement funds are obviously important too, but since I can’t access those until a certain age, I track them separately.

    There will always be a problem when trying to estimate the FMV of any assets for net worth purposes (even aside from the ridiculous volatility of the housing market lately).

    Paul

  25. If you think the price estimates are inaccurate, then by all means stop using them. But just because they are volatile does not make them inaccurate, it might just mean that we are currently in a volatile market.

  26. What about the appraised value provided by your local property (tax) appraiser’s office (I guess the name and applicability vary by state), or is that not applicable or accurate?

  27. I think this experiment proves what most of us already know…these estimates are very poor. Zillow was off by 20% on the value of my house immediately after I bought it. If it’s not smart enough to use the purchase price of my house to determine the value of my house, what does it say about its system? Last time I checked it said my house was worth between $210,000 and $340,000. That’s some precise calculating.

    So I agree with the first poster. You can say you have personal bias, but I would trust you and your bias with your human brain and human eyes over some idiotic program. Find 3 houses similar to yours and see what they sold for. That’s the value of your house more or less, and no program is going to know anything more than that.

  28. Debt Free Dude says:

    Even if they are pulling from the same databases they may be updating at different times. Current house prices are fluctuating quite a bit in some areas and that is going to cause there to be quite a wide spread in their pricing.

  29. For me, figuring out “net worth” isn’t about how much I’m worth to my heirs, but to measure how close I am to my goal of financial independence, and to track how well I’m on track to making that goal. Therefore, I don’t count in things like personal property or my home. But this is easy for me, because I didn’t start keeping track until after we bought the house.

    When you first were looking at buying a house, and you were asking about whether to include the house in your net worth figures, I was one of the ones who said you should include it – you were putting a sizable down payment on the house, and I thought it would skew your net worth history too much to just have that amount disappear. But like you, I think, I’ve not been happy with the way things have jumped around in your figures because of the house.

    Now, I don’t know what the answer is. If it were me, I’d try to come up with some way to gradually work the house out of my figures – figure some sort of “paying off” not your mortgage, but the down payment amount. Does this make any sense at all?

  30. Depends why you want to know your net worth. I only care about my spending needs now and in retirement.

    I exclude both my mortgage liability and my home value (asset) from my net worth estimate.

    I have calculated my investment and retirement goals on the assumption that I have a home to live in, and my mortgage payment is my “rent”. At some point (mortgage paid off) my “rent” drops to cost of property tax only.

  31. So you are aware:
    Zillow has been making great strides to make their Zestimates accurate. The problem is that they do not have access to all of the data (Realtor Associations make quite a bit of data on their MLS proprietary- which prevents accurate appraisals by others).

    Coldwell Banker would have access to that data, but there could be a number of factors for their swings. If a few foreclosures are thrown into the mix, it would suggest all home values are decreasing when they might not be. If one valuable home sold, the opposite would occur. I wonder if their estimator takes in appraisal figures? If so, this year has been a wild ride for appraisers, who have also been denied access to MLS data in an effort to make values more fair. The problem which follows from this situation is an appraiser in California will never see a home in New York, but they will place a value on it.

    From experience, I would say stick to Zillow. It is not entirely accurate, but they are getting closer on average to any other system.

  32. MyCrossoverPoint: I’m estimating my rent based on what I see in the local paper and pennysavers for home rentals in my area, and online under my local MLS rental website. I’ve actually been trying to get an idea of value in the event that I end up renting my house in the next few years, when I buy a new home. But now, actually, I’m toying with the idea of renting out my home, and then renting a home for myself in the best school district in the area. It may cost me about $600 more a month/net to rent a home in the best area, but I won’t have to commit to a purchase, and I won’t have to fix stuff or buy stuff for the house — I can be a renter. It’s funny, I always liked the idea of owning my home, but the money I pour into this thing! Renters can be really lucky, they get to experience the home without any of the headaches. I wouldn’t want to rent an apartment in someone’s house, but the whole house would probably be fine. Though, we’ll probably end up buying, because I don’t think my husband would be as happy.

  33. These automated valuation models are limited in accuracy – they are just estimates. With that said, I think this year has seen real volatitly, largely to the upside and Zillow’s model is closes to the blended average for the group – because they probably have perfected their AVM model the furthest. But it is limited by lack of MLS data and I wouldn’t necessarily assume that Coldwell Banker has a full set of accurate sold data on a national basis, though it may theoretically have access!

  34. I simply use the assessed value (for property tax purposes) minus 6% (to reflect theoretical realtor fees). I update annually when the new assessments come out. That’s accurate enough for me.

Trackbacks

  1. Are Zestimates Really Accurate? | Real Savvy Real Estate says:

    [...] smaller, study came to a similar conclusion. The personal finance blogger behind MyMoneyBlog.com tracked his home’s value over a one year period, using several different online valuation tools. Here [...]

  2. [...] Value I am no longer using any internet home valuation tools to track home value. After using them for a year, I went back to simply taking a conservative estimate and focusing on mortgage payoff. After [...]

Speak Your Mind

*