One More Look at Bogus Reports from the NWMLS

I’ve got one last set of data that I’d like to share on the matter of the NWMLS monthly reports not really reporting what happened during the specified month.

There’s a real estate statistics service called Trendgraphix that puts out monthly reports as well. Since they are plugged directly into the MLS database, they are able to report on the same underlying data as the NWMLS monthly reports. From what I can tell, Trendgraphix does their best to only report what actually happened during a given month. For late-reported sales, they will update the numbers from previous months when they publish each monthly report.

I had a friend with Trendgraphix access go in and pull the last few years of King County closed sales counts for me. Here is what it looks like when you compare the Trendgraphix reports (which are pulling directly from the NWMLS’ own database) to the NWMLS reports:

Trendgraphix and NWMLS Reports

There has been an ongoing difference between the two for as far back as this data goes, but July 2010 saw far and away the largest variance.

While it may be true that in the long run, the over and under-reporting in the NWMLS reports wash out, that doesn’t change the fact that they are knowingly publishing false information when they claim that there were X home sales closed in a given month, when X includes closings that actually occurred in previous months.

If you just want to look at a long-term chart and get a feel for general trends, the NWMLS monthly reports are fine. However, if you’re trying to get an idea of what is really going on in a specific given month, it would appear that you should look elsewhere.

Previously:

  

About The Tim

Tim Ellis is the founder of Seattle Bubble. His background in engineering and computer / internet technology, a fondness of data-based analysis of problems, and an addiction to spreadsheets all influence his perspective on the Seattle-area real estate market.

23 comments:

  1. 1
    hp says:

    So this makes sense.. If you just look at last 6 months and add up all the numbers
    -47-90-92-22+49+215 = +13 which I am sure will wash out in the next month or so.. But agree, they still are knowingly reporting incorrect data or they just need to change the legend to say “Sales reported in July” rather than “Sales happened in July”.

    Rate this comment: Thumb up 0

  2. 2

    RE: hp @ 1 – They seemingly just say “sales.” I would agree “sales reported” would be better if that is in fact what is going on (that it’s not more complicated than that).

    Using the graph above, seemingly under-reporting sales by a significant margin is more of a problem than over-reporting them. I don’t think those numbers in the graph above exactly match what I came up with (I lost my notes and don’t want to re-do the work), but using those number only two months prior to July 2010 had over-reporting discrepancies above 50. 5 months under-reported by more than 50.

    Rate this comment: Thumb up 0

  3. 3
    JChristopher says:

    Tim,

    It looks like a wash, but it makes sense. They would want to understate the numbers before the tax credit expired, and then overstate them once it does expire. This way they can reinforce the illusion that the economy is okay, and all these socialist programs aren’t what we’re hanging on by a thread to.

    Rate this comment: Thumb up 0

  4. 4
    2kt says:

    The feds deal with this issue by releasing adjusted numbers the following month, to account for slow/late reporting and allow for adustments.

    The NWMLS just lumps it all together and reports one time. The solution is to hold a release till mid-month and give harsh penalties to those realtors who report sales late.

    In any case, calling this type release “bogus” amounts to creating a storm in a glass of water, dear esteemed proprietor.

    Rate this comment: Thumb up 0

  5. 5
    per_se says:

    I don’t necessarily agree with your position on this one Tim. Realistically, they should offer updated stats like Trendgraphix or how gov stats are provided and then updated as more reliable information comes in.

    I think you need to look at expediency vs absolute accuracy. What is the avg variance between the month and sales reported? If it takes up to two months to have all the sales for a given month reported to have at least 95% of sales vs reporting a couple days after the end of the month and, I’m partially guessing based on the graph, a 5-8% discrepancy. I’ll take the expedited report.

    They should make it more clear on what they are reporting, with this and everything else, but I don’t think it’s wrong of them to report the sales numbers for a previous month like that.

    Rate this comment: Thumb up 0

  6. 6

    The easiest fix would be to have the NWMLS prohibit firms from handing out commission checks until the status was updated. ;-)

    Rate this comment: Thumb up 0

  7. 7
    LN e says:

    If the inflation in the June and July NWMLS numbers are the result of prior-month late reporting, todays chart would suggest that many of the late reports are more than 5 months over due. So the addition of Matrix in May doesn’t justify all the lates. IMHO the average RE agent should have a lot of time to work on filing reports in light of home sale volumes relative to historical norms.

    Rate this comment: Thumb up 0

  8. 8

    By LN e @ 6:

    If the inflation in the June and July NWMLS numbers are the result of prior-month late reporting, todays chart would suggest that many of the late reports are more than 5 months over due. So the addition of Matrix in May doesn’t justify all the lates. IMHO the average RE agent should have a lot of time to work on filing reports in light of home sale volumes relative to historical norms.

    I have no idea why you think it would suggest that, but in any case it’s not correct. Virtually all of the reported late in July (almost 200) are from the prior three months, with the bulk of those (over 100) from the June.

    Vague numbers from NWMLS sources, but not compiled or guaranteed by the NWMLS.

    Rate this comment: Thumb up 0

  9. 9
    buystocks says:

    Do you happen to know what the median reported by Trendgraphix was for July?

    Rate this comment: Thumb up 0

  10. 10
    The Tim says:

    RE: buystocks @ 9

    Trendgraphix rounds their reported median to the nearest $1,000. Here’s the last year or so:

    Month  | NWMLS    | Trendgraphix
    May-09 | $375,000 | $380,000
    Jun-09 | $395,000 | $396,000
    Jul-09 | $384,000 | $380,000
    Aug-09 | $375,000 | $376,000
    Sep-09 | $382,160 | $385,000
    Oct-09 | $377,500 | $374,000
    Nov-09 | $370,000 | $378,000
    Dec-09 | $380,000 | $380,000
    Jan-10 | $375,000 | $372,000
    Feb-10 | $373,010 | $370,000
    Mar-10 | $367,250 | $370,000
    Apr-10 | $375,000 | $375,000
    May-10 | $379,000 | $380,000
    Jun-10 | $383,000 | $385,000
    Jul-10 | $399,950 | $400,000

    Rate this comment: Thumb up 0

  11. 11
    deejayoh says:

    If the NWMLS really cared about the statistics, they wouldn’t report the median as the best indicator of market trends. They could easily report paired sales, or $/sqft trends. But they don’t. Because they aren’t really all that sophisticated in how they report things or concerned about the 3rd decimal point accuracy of what they report. And really, I think we all knew that already.

    Rate this comment: Thumb up 0

  12. 12
    David Losh says:

    The statistics are only used for sales talk. The median price going up will be used to generate sales. It can be used to show the market is stable.

    What I keep wondering is that the statistics kept saying buy, buy, buy in 2005, 2006, 2007 and contrary to that you started a blog claiming there was a bubble. I think you were right.

    Now it’s every month with the statistics like they have some meaning, which they don’t. They are only talking points for Real Estate sales people.

    Where’s the market going? or is that one of those crystal ball kinds of questions?

    Rate this comment: Thumb up 0

  13. 13

    RE: deejayoh @ 11 – Paired sale and square foot trend can be just as inaccurate. The reason there’s a problem here as to median is because there was a huge artificial stimulus that affected the market. If you suddenly had an annual tax on bathrooms, and then removed it again, median, paired sale and square foot price would all be affected significantly in short order.

    Rate this comment: Thumb up 0

  14. 14

    RE: The Tim @ 10 – I find it very hard to believe that the median without those 200 sales would only be $400,000. For that to be the case there would have had to have been a ton of transactions at between $395,000 and $400,000, because the median on the late reported sales was much lower than $400,000. Their median was coincidentally closer to the median of their period.

    Rate this comment: Thumb up 0

  15. 15
    patient says:

    Great, the only nwmls measure left with some value is now showing to be bogus as well…
    Thanks The Tim for putting in the effort to report and investigate these things.

    Rate this comment: Thumb up 0

  16. 16

    RE: deejayoh @ 11
    Price per sq ft would also suffer from not reflecting the mix of sales. Larger homes tend to sell for significantly less per square foot. Paired sales would be interesting to look at.

    Rate this comment: Thumb up 0

  17. 17
    deejayoh says:

    By Kary L. Krismer @ 13:

    RE: deejayoh @ 11 – Paired sale and square foot trend can be just as inaccurate. The reason there’s a problem here as to median is because there was a huge artificial stimulus that affected the market. If you suddenly had an annual tax on bathrooms, and then removed it again, median, paired sale and square foot price would all be affected significantly in short order.

    Perhaps I should have said “best and only”. I agree there can be issues with any measure – but the point is they have the data to report any way that they choose. And they could choose to be more transparent and accurate, but they don’t

    Rate this comment: Thumb up 0

  18. 18

    By Ira Sacharoff @ 16:

    RE: deejayoh @ 11
    Price per sq ft would also suffer from not reflecting the mix of sales. Larger homes tend to sell for significantly less per square foot. Paired sales would be interesting to look at.

    But paired sales suffer from changes in condition, and/or remodels.

    Rate this comment: Thumb up 0

  19. 19

    By Kary L. Krismer @ 18:

    By Ira Sacharoff @ 16:
    RE: deejayoh @ 11
    Price per sq ft would also suffer from not reflecting the mix of sales. Larger homes tend to sell for significantly less per square foot. Paired sales would be interesting to look at.

    But paired sales suffer from changes in condition, and/or remodels.

    So we’re left with……..nothing?

    Rate this comment: Thumb up 0

  20. 20

    RE: Ira Sacharoff @ 19 – All these things are just an indication of market strength. Right now we know the median increase is over-estimating market strength, in part because another indicator (volume) is down. Over time there will be changes that through it off. Very long term it’s larger houses and more bathrooms (overstated strength). More recently it was the addition of short sales, REOs, and a first time buyer credit (overstated weakness). This month it’s the elimination of a lot of tax credit (overstated strength). At least this time we know what’s causing the change. That hasn’t been so apparent with the addition of short sales and REOs.

    But in any case, we’re not left with nothing. There are always appraisals and BPOs that look at specific properties.

    Rate this comment: Thumb up 0

  21. 21
    Greg Perry says:

    I’ve used Tredgraphix for years and have often said their corrected data is more accurate.

    Rate this comment: Thumb up 0

  22. 22

    […] release the stats so early in the month, but maybe they figured since the cat is out of the bag on their bogus methodology already anyway, they would go ahead and push the report out as early as possible. Here’s what […]

    Rate this comment: Thumb up 0

  23. 23

    […] for each of our eight covered counties as of September 2010.CAUTIONNWMLS monthly reports include an undisclosed and varying number ofsales from previous months in their pending and closed sales statistics..CNNTable {margin: 5px […]

    Rate this comment: Thumb up 0

Leave a Reply

Use your email address to sign up with Gravatar for a custom avatar.
Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Please read the rules before posting a comment.