Misleading NWMLS Stats Hide Severity of Sales Dropoff

June market stats have been published by the NWMLS. Here’s a link to the NWMLS press release (not yet live as of this posting).

As requested, here are the full monthly stats from the NWMLS “July” report. Before we get into the numbers though, there are two key points that I’d like to try to clarify.

First, thanks to some in-system sleuthing by Seattle Bubble regular Kary L. Krismer, we believe we have discovered the explanation for July’s 200+ magical mystery sales. Basically it appears that when the NWMLS puts out their monthly reports, the sales counts are not really the number of sales that happened in the reported month. These counts also include sales that took place in previous months, but were updated in the system by member agents during the reported month.

In other words, the extra 200+ sales in July’s report were actually sales that took place in June (or possibly further back). Kary explains:

In 2009 the average number of late reported transactions each month was about 110-115. April, June and July 2010 saw significant increases, and were all higher than any month in 2009. Those late sales have to be reported sometime, and in just one week in July over 100 late sales from June were reported. That one week would have been pretty close to an entire typical month for all prior months in 2009.

As for the reason for the increase, again one of my guesses is Matrix. It kicked in during May, and the beginning of May is when many of the sales from April should have been reported (due to how many sales close at the end of a month). April had twice as many sales reported late as March.

So anyway, it appears much of the problem is due to agents, and specifically agents using the new Matrix system. Remember that many agents might not have a sale for a two month period, so three months into this there are some agents who still have not reported a sale, and have not learned how to do it.

Numbers from NWMLS sources but not complied or guaranteed by the NWMLS.

At least we have solved the mystery. Unfortunately this means that the monthly stats from the NWMLS are much less useful than I had believed them to be. They’re still decent for showing general trends, but if you really want to know what actually happened just in July, you’ll have to look elsewhere.

The second point I’d like to spend a little time on is the median price. We have covered this topic in a fair degree of detail here in the past, and I highly recommend you read the first portion of this post for a quick refresher.

The important thing to remember when looking at July’s median price data is that between June and July a large portion of the low end of the market basically vanished. Since sales distribution is an inherent factor that drives the median price, the expected result of a disappearing low end would be an increase in the median. Incidentally, that happens to be exactly what happened, as King County’s SFH median jumped up 4.4% in a single month.

Unfortunately, this bump in the median only really tells us what we already knew: that sales of cheaper homes fell off in July. What it does not tell us is that the actual value of individual homes is actually on the rise. We can say “the median price of houses sold in King County was up 4% in July from a year earlier,” but that basically means squat when it comes to home values.

So, with those two points in mind, hit the jump if you’d like to check out this month’s misleading NWMLS data.


NWMLS monthly reports include an undisclosed and varying number of
sales from previous months in their pending and closed sales statistics.

Here’s your King County SFH summary, with the arrows to show whether the year-over-year direction of each indicator is favorable or unfavorable news for buyers and sellers (green = favorable, red = unfavorable):

July 2010 Number MOM YOY Buyers Sellers
Active Listings 10,470 +6.0% +6.2%
Closed Sales 1,474 -21.6% -14.6%
SAAS (?) 2.31 +37.9% +15.3%
Pending Sales 1,716 -4.9% -22.6%
Months of Supply 5.47 +3.1% +38.7%
Median Price* $399,950 +4.4% +4.2%

Feel free to download the updated Seattle Bubble Spreadsheet (Excel 2003 format), but keep in mind the caution above.

Here’s your closed sales yearly comparison chart:

King County SFH Closed Sales

We got the drop we were expecting to see, but the messed-up way the NWMLS reports “monthly” statistics hid the true severity of the fall.

Here’s the graph of inventory with each year overlaid on the same chart.

King County SFH Inventory

Woo, 10,000. I wonder if we’ll manage to break 11,000 this year.

Here’s the supply/demand YOY graph. In place of the now-unreliable measure of pending sales, the “demand” in this chart is represented by closed sales, which have had a consistent definition throughout the decade.

King County Supply vs Demand % Change YOY

Despite the NWMLS monkeying with the data, the sales and listings YOY lines crossed back into strong buyer’s market territory in July.

Here’s the median home price YOY change graph:

King County SFH YOY Price Change

See the note above about what this is really telling us.

And lastly, here is the chart comparing King County SFH prices each month for every year back to 1994.

King County SFH Prices

Even still, the median barely edged above 2005 levels.

Tomorrow we’ll check out all the local news reports and see if any of them managed to find the real story this month.

0.00 avg. rating (0% score) - 0 votes

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. Tim also hosts the weekly improv comedy sci-fi podcast Dispatches from the Multiverse.


  1. 1

    It’s Still Convoluted and Subpar IMO

    Parking your hat on a “verbal hatrack” explanation for the data anomalies Tim described is like throwing your hat in the air and expecting it to land on something solid, besides the floor. In layman’s terms, it won’t hold up in court.

    I use Microsoft Excel to get statistical analysis adequacy [acceptable correlation factors of experimental data] trending for medical research reporting. If the correlation factor is inadequate, the results of the test are inadequate to be published. Bill Gates provides me with the correlation factor of the data I’m trending on Excel. Legally, I’m off the hook, the Excel S/W does the imbedded math for me [as long as I reference Excel as the correlation factor source, which I do].

    Now, in Tim’s test data referenced above, what the Hades is the “correlation factor” that proves mathematically and legally in writing, that monthly missing data makes the monthly data acceptable to publish as usable? LOL, you come up nada don’t you.


  2. 2

    RE: softwarengineer @ 1

    My Advice to Prove Out the NWMLS Monthly Data Correlation Factor

    First, you need to get a range on data that’s missing and/or distorted [LOL, good luck], then do a trend analysis thesis paper on doing the test results with best to worse scenarios and get a “correlation factor” that matches NWMLS data quality control using the “range”.

    If you hired me for a substantial fee, I could solve this dilemma [I don’t come cheap]…..but my guess is NWMLS doesn’t even have a correlation factor QC limit, so it would need defining too….LOL

  3. 3
    Drshort says:

    Re: mix shift and the high median

    I dont recall a big discussion in early 2009 that the lower median was caused by a collapse in jumbo financing.

  4. 4

    First, I think it was Dr. Short that provided the key information that lead to my conclusions on the data. He indicated a type of search I’d not done before.

    Second, at this point, unless a different explanation is given, I think the best way to view these reports is that they represent sales reported during a given month, not sales occurring during a given month. The vast bulk of the sales will be from the month, but some will not be.

    Third, assuming my analysis of what’s happening is correct though, at this point, however, it’s premature to say that the reported number overstates actual sales. We won’t know that for a couple 2-3 months (well actually non-agents will never know it). Over the long run the average sales reported from prior months would equal the sales not reported in a month.

    Fourth, assuming the sales not reported yet for July are anywhere near the sales reported from prior months, it’s very likely that the actual median for July was higher than reported. Stated differently, the effect of the tax credit expiring was larger. Also, I would note that for the Seattle only data, the YOY increase was 10% ($400,000 to $440,000). I haven’t done a separate analysis of Seattle, but in spot checking the areas it did appear some were way off in reporting and some almost spot on. It’s possible Seattle agents were better at reporting, and that the 10% number better reflects the tax credit change. Just a theory.

  5. 5
    Jeremy says:

    Isn’t the inexactness of the data baked into every monthly report from the MLS though? So I would think that it would make the historical data comparable but readers should just be aware that the numbers aren’t precise but in the long run are accurate excepting the timing? I agree that 100% accurate data is nice but don’t know of a good source for that in any large dataset other than looking at the public records of closed sales. And I’d even bet there are errors there too but probably far fewer because people have a financial incentive to get the transactions correct.

  6. 6
    Cheap South says:

    By now, this site is followed to a certain degree by the local industry and its reporters. Hopefully, someone will consider the explanations before they typed “Homes up over 4%; it’s now or be priced forever!!!”

  7. 7

    RE: Jeremy @ 5 – As I mentioned before, it’s speed vs. accuracy. If the NWMLS reported as late as Case-Shiller, their numbers would be much more accurate, but we wouldn’t get the information for months, which would make it less useful.

    Still though I think the NWMLS could do a bit more. For example, an automated email to agents with pending listings on the first day of the month, reminding them to update any listings that sold.

  8. 8
    Dave0 says:

    Maybe each month needs to have a number of “revisions”, like when the government is reporting GDP. This month can have the preliminary July data, along with the second revision of June data (adding in some sales that were reported in July), and maybe a third revision to May data (if any sales were reported 2 months late). However, this might require access to the NWMLS and compiling your own stats from within the system rather than relying on what the NWMLS reports.

  9. 9

    RE: Dave0 @ 8

    Government Data Reporting Anomalies Worse

    GDP has always been disputed for it’s inacuracies by a plethora of professionals [the govenment too], but when they find them [and they do, big discrepancies too, like 10s of millions the BLS unemployment hoards they ignored in 2009/2010 and admitted it too], are they corrected later in like, the real unemployment rate?….Hades no….LOL

    God forbid we paint a bad picture even worse. I.e., GDP ignores the costs to get our crumbling infrastructures and wore out process tools replaced….just a minor over-sight…LOL

  10. 10

    As to the yellow “caution” warning above, I think we’ve always known that pending data includes prior months, or at least that there’s no way to verify any different date than when reported by the agent. With sale transaction you can look to the recording date of the deed. With the pending date the information is private to the parties.

    And again, I think we pretty well suspected this with sold data. Am I the only one that remembers the discussion of how the NWMLS reports are never updated, and that they have to have some way of picking up the late reported sales? With the YTD data always equaling the sum of the prior months’ reports, it pretty much had to be that way.

  11. 11
    dw says:

    Given that there were a lot of sales pushing through the system with the stimulus, I can see how this current circumstance is unusual.

    I’m with Kary, though — if this is how they’ve always done it, there’s nothing inaccurate about it, so long as the reporting is consistent. It may be mislabeled and hyped, but if it is consistent, it’s a valid measure.

    And I’m not a real estate hype man, but I think the indicators are suggesting we may finally be within sight of bottom. I might be wrong and that light at the end of the tunnel could be the double-dip recession or a wave of deflation, but it does look like the price fluctuations are settling closer to 0%.

  12. 12
    dw says:

    And getting better data in the hands of the people is the mantra of the Open Government movement. I think you can already access some of this data (and most of it is public record through KingCo anyway), but an API hanging off the databases would make it much easier.

    Too bad the people of this state don’t want to fund stuff like that.

  13. 13
    Flotown says:

    The price/sf in North Seattle seems to be falling substantially. Most sales over the past two weeks have been in the $220-$260 range, while the average stats on redfin show the average to be in the $300+ range.

  14. 14

    RE: Flotown @ 13 – I’m not going to run those numbers but my guess would be with the end of the tax credit the properties being purchased are larger square footage, and larger square footage results in lower price per square foot.

  15. 15
    GH says:

    Sorry, Tim, I love your website but do not agree:

    (1) Is there any evidence that home sales are higher in the upper price ranges versus the lower price ranges? I’m not aware of any studies that say that.

    (2) The median price increase is consistent with a long-term stabilization trend that we’ve seen over the last 6 months. The Case-Shiller Index confirms this.

    The weight of evidence suggests that the Seattle market is stabilizing and improving.

  16. 16

    RE: GH @ 15 – I seldom make predictions, but even I was expecting the median to rise as a result of the tax credit going away. That’s what happens when you remove what is largely below median sales from the mix. We’d been talking about that possibility for a couple of months or so.

    If the market had seen some severe downturn it might have overpowered that effect and changed that result, but otherwise an increase was pretty well expected.

  17. 17
    per_se says:

    RE: GH @ 15

    You can’t look at the median price alone. A rise in inventory (supply), a drop in sales (demand) and all after the conclusion of a tax credit where anyone making over $125k/$225 household was ineligible would be consistent with a change in the buying mix,not a stabilization.

  18. 18
    Sniglet says:

    RE: GH @ 15

    The weight of evidence suggests that the Seattle market is stabilizing and improving.

    Watching the local housing market trends, or statistics, can cause you to lose perspective. The direction of the Dow Jones stock index matters FAR more to prices of homes in the Puget Sound than tax credits, available inventory, or anything else. If you believe that stock markets are stabilizing, then it would be reasonable to assume that a similar thing would happen for housing. Unfortunately, I happen to believe that the real pain in stock valuations is yet to hit us. If this is the case (i.e. that the stock indexes will fall significantly), then housing prices will do likewise.

    Just think of it this way: how can housing prices “stabilize” if unemployment increases (which is what would certainly happen with a crashing stock market, which indicates a contracting economy)?

  19. 19
    Ahau says:

    I’m with the others that aren’t that upset by the over reporting in July due to late reportings in the MLS. Long term, things should wash out as long as people are being consistent in their reporting. This same anomaly would have made sales numbers look lower at the run up from the tax credit, but no one was jumping over the MLS for under reporting the sales numbers at that time, iirc.

    I’ll apply the same logic to the median sales price–the tax credit likely artificially lowered the median sales price vs actual home values while it was in effect, by shifting the mix towards cheaper homes (though this was tempered by the actual but temporary increase in value that it created). It will be interesting to see how things normalize in coming months, I have no clue which direction things are heading, but the pending sales vs. inventory graph isn’t encouraging for the bulls!

  20. 20
    Cara says:

    By GH @ 15:

    Sorry, Tim, I love your website but do not agree:

    (1) Is there any evidence that home sales are higher in the upper price ranges versus the lower price ranges? I’m not aware of any studies that say that.

    Hopefully the Tim will soon do his post where he breaks down the sales into price brackets to see the distribution of sales. If you’re skeptical that the $8k pulled forward low-end demand and purchases but not high end, thereby skewing the distribution and raising the median, just wait until that graph comes out. You’re right that the median moving in the direction we all expected in and of itself doesn’t our theory is what happened.

    It’s fair to be skeptical.

  21. 21
    deejayoh says:

    By GH @ 15:

    (1) Is there any evidence that home sales are higher in the upper price ranges versus the lower price ranges? I’m not aware of any studies that say that.

    Data supporting this was provided in this Seattle Times article. One of the few useful bits in there.

    In Seattle, for instance, 40 percent of the houses that closed in July sold for $500,000 or more, compared with just 26 percent in the same month last year, Windermere Real Estate spokeswoman Sonja Riveland said in an e-mail.

    you could also pull up sales in Redfin by price bucket to validate this

  22. 22
    hoary says:


    I would contact the county records office for aggregate data. That way you’ll know the date a property is sold, and also the date the property is recorded. Many times there’s spillover from month to month, sometimes much longer.


  23. 23
    David Losh says:

    RE: Sniglet @ 18

    The way I have always looked at it is, that the stock market reflects the Real Estate Industrial Complex.

    Microsoft changed the way we look at intellectual property, but we still call it property. Business Opportunity is a part of a Real Estate.

    When the economy is good, or flowing, business buys more equipment, hires more people, and develops more property. The stock though, even with all of it’s returns on invested dollars, still wants to see tangible assets, as proof there is some skin in the game.

    If I build my business with rented equipment, leased space, and temporary labor, even the promise of full time employment, investors will want to see something to secure the investment dollars with.

    To me it’s all round, and round.

  24. 24
    Sniglet says:

    RE: David Losh @ 23

    To me it’s all round, and round.

    Agreed. It is highly unlikely we will see a de-coupling, where one part of the economy does well while other parts do poorly. Housing isn’t going to boom if there aren’t jobs, and stock markets won’t shoot up while house prices crash either.

    To believe that stocks or housing prices are “stabilizing” is to believe that they are ALL stabilizing.

  25. 25
    Scotsman says:

    RE: Sniglet @ 24

    There won’t be any decoupling. We need to revisit the basics- home prices are driven by incomes, financing, and aggregate demand/supply relationships. Last I checked wages weren’t headed up, the number of unemployed hadn’t changed significantly, there wasn’t any easy financing available and we still had almost eighteen million vacant housing units in the country.

    It’s a pause, not a bottom. And if it were somehow a bottom then we will be staying here for some time until employment and income numbers start to move up, supporting higher housing prices. Understanding isn’t hard, but accepting reality and change is.

    The next 20% down will hurt just as many people as the last drop. Don’t be fooled by the machinations of those in power during an election year. The whole facade is held together by the faintest of illusions, nothing more.

  26. 26

    By Sniglet @ 24:

    RE: David Losh @ 23

    To me it�s all round, and round.

    Agreed. It is highly unlikely we will see a de-coupling, where one part of the economy does well while other parts do poorly. Housing isn’t going to boom if there aren’t jobs, and stock markets won’t shoot up while house prices crash either.

    To believe that stocks or housing prices are “stabilizing” is to believe that they are ALL stabilizing.

    Completely agree, but housing, jobs, and the stock market are seldom marching in lockstep with each other, so sometimes people can get the impression that things are stabilizing.
    For example, the stock market saw a climb from 6500 to about 11,000 before it slipped a bit. Does this mean that jobs and housing will follow and we’ll see a big increase in home sales, home prices, and lower unemployment? It sure doesn’t look like that.
    In 05-06 it was obvious to me that home prices were going to take a dive. Sales had leveled off, and inventory was huge. But even then things just sailed along inexplicably for a couple of years. A rising stock market could make housing prices fall more, if it makes it look like stocks are safer, more of a sure thing, etc. But it also looks like stocks are still pretty rich even after falling a little, and have more to fall.

  27. 27
    David Losh says:

    RE: Sniglet @ 24

    One of the directions Real Estate will take is in wholesaling. I’ve noticed ads for wholesaling brokers, or people who will take multiple properties from a bank.

    There was some legislation about this in 1988 after the Savings, and Loan scandal, but I think this will be the newer trend. I also think that there will be more give, and take between stocks, and Real Estate, manufacturing plants, and factories. Let’s not forget agricultural, and tree farming.

    Stability, in my opinion, will be where safe returns are found for a better annual rate of say 6%, maybe 9%.

    I think the days of 17% Fidelity Magelin are over, but you never know.

  28. 28
    Hugh Dominic says:

    RE: Ira Sacharoff @ 26 – rising earnings, and rising stock prices can go in step with rising unemployment. That’s what you get when technology advances are harvested by the companies as productivity gains.

    That’s what’s going on right now. Companies are shedding jobs because it just doesn’t take so many people to do the same stuff they did five years ago. Also see: The Internet Took Your Job.

    This has one of two endings:

    The happy ending, where stable and passive government policy allows the free market to complete a short but scary adjustment, which ends when the economy discovers innovations that employ people in new jobs that they had not even conceived of today.

    The Obama ending, where government protection for rotting, dead industries and job skills lock the USA in an obsolete economy, drive up a huge debt load, and prevent a recovery of the middle class.

  29. 29
    mid says:

    Don’t have a meaningful contribution here – but this must be why Redfin values have gone up all crazy! At least in the Bellevue area. It certainly doesn’t seem to be reflected in the sales of homes in the area.

  30. 30

    RE: mid @ 29

    Perhaps It Was Money Magazine Recently Making Bellevue One of the Most Livable Cities in America

    LOL, with a $500K avg home price….livable for who? The hoards of new higher incomes MSFT and Boeing are hiring? LOL

    A big problem with even counting upper middle incomes in Bellevue, are the actuals too….most of these folks are up to their eyeballs in debt, most already own/live in like $500K homes…LOL

    No, we need fresh $150-250K household incomes for Bellevue real estate to survive, last I heard, these were being butcher axed out.


  31. 31
    ray pepper says:

    just got this “typical” email from a client this morning………..

    “just a heads up for ya…cant buy a home right now…got laid off yesterday…lame…”

    housing rebound?……………not….
    flat with a trend line down bias for the next decade…………absolutely

    They r all coming back…but never discount the FED and further stimulus to cushion

  32. 32
    Magnolia44 says:

    If we see another 20% down jingle keys will be a guarantee for this household. Squatting will recoup a substantial part of my down payment and they can come after us all they want for any kind of deficent second. My 827 credit score will be kissed goodbye but what’s credit when you live mostly by cash anyway.

    Just for good measure I would go ahead and replace the older car with a fresh set of wheels before doing so since monthly cash flow will be freed up substantially.
    I am already ok with jingle keys, I think. Its the wife who would be harder to convince but the additional leg down would assure her its the right move.
    Feel bad for neighbors? Nah… One was willed his home, other one lives for free in Mommys home and other paid half what we did they will all be ok.

    Cheers all!

  33. 33
    Cheap South says:

    RE: Magnolia44 @ 32

    If it happens it will take about 24 months.

  34. 34

    […] appear that you should look elsewhere.Previously:Magical Mystery Sales Appear in Latest NWMLS DataMisleading NWMLS Stats Hide Severity of Sales DropoffPosted in Statistics | Tagged NWMLS, Trendgraphix Posted by: The Tim Tim Ellis is the founder of […]

  35. 35

    […] you know, I regularly criticize the local multiple listing service for their lousy accounting methods that result in frequent blatant mis-statements of basic facts. Now multiply these types of issues […]

  36. 36

    […] data download puts late-reported sales into the month that the sale actually took place rather than in the month they were reported, there is a slight difference in the number of sales I’m counting vs. what the NWMLS reports […]

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=""> <s> <strike> <strong>

Please read the rules before posting a comment.