Commentary by Tom Ruff

As an analyst, the number one question I get asked is, “What percentage of homes sold are listed on the MLS?” This month in STAT, we’ll attempt to answer this question as well as place this answer in an historical perspective. We’ll also consider the value aspect of these sales, comparing the value of homes sold on the MLS to homes sold off market.

The answer to this question, like every other question I’m asked, is… it depends. It depends on how you define your data. For this discussion, we’re going to look at all homes sold in Maricopa and Pinal counties as defined by public record and eliminate Trustee Sales (trustee sales are homes sold at foreclosure auction on the court house steps). There have been 1,584 foreclosure auctions held in Maricopa County this year. Of the properties sold at auction, 602 reverted to the beneficiary and 982 were purchased by a third party. To put these numbers in perspective, on average only 5.78 homes were purchased daily at auction for the 170 business days of this year, and only 3.54 reverted to the lender/bank.

When we do a down and dirty comparison for the first 7 months of this year, we reported 57,034 home sales. Public records tell us 76,200 homes sold for this same period in Maricopa and Pinal counties. Using this simplified methodology, the percentage of homes sold on the MLS is 74.84%.

So, the answer is 74.84% of homes are sold on the MLS? Again, it depends. I think what we’re really trying to define is the percentage of resale homes listed on the MLS. For the first 7 months of 2017 there have been 10,287 newly built homes sold, but only 26.28% of new constructions are listed on the MLS. When we remove newly built homes we show 65,844 sales. In other words, 81.44% of all resale home sales are listed on the MLS.

Does this mean 18.56% of the homes sold this year are FSBO (For Sale By Owner)? No, not at all. When we think of FSBOs, we traditionally picture a homeowner that sells their home without using an agent. We might also picture an FSBO sign in the front yard, a Craig’s list ad or Zillow’s Make Me Move or Instant Offers. I believe it’s impossible to identify a true FSBO from our existing data sets. Our only reliable conclusion is that 18.64% of all publicly recorded resale homes are not listed on the MLS. Today we have “iBuyers” like Opendoor and Offerpad and for over 20 years we’ve had We Buy Ugly Houses. We also have wholesalers and a myriad of small investors that market directly to the homeowner. Wholesalers and investors love the “D” sales: death, divorce and distress. All of the above purchases are a result of direct solicitation from the buyer. Since our industry loves acronyms so much, maybe we need a new one like BOSA (Buyer Offered Solicitation Accepted). On the purchase side, none of the BOSAs are listed on the MLS. But after the properties have been purchased and flipped, our data tells us 75% are listed on the MLS. The “iBuyers” list nearly all the properties they’ve purchased on the MLS.

Let’s take all the homes purchased in Maricopa and Pinal counties in the first 7 months that were not listed on the MLS. If we look for buyers’ names or addresses that appear multiple times, we can identify parties that directly market to potential sellers. The one metric we cannot account for are “pocket listings.” We know there can be a variance in the percentage of properties listed on the MLS and this variance is directly tied to market conditions. Low supply and high demand increases the likelihood of pocket listings while high supply and low demand has the opposite effect. The stronger the sellers’ market, the greater likelihood of pocket listings. At present, we do not have a means of identifying pocket listings. Since we can’t clearly define a true FSBO from our data, and we can’t identify pocket listings, we’ve combined the two in the table below.

Type Sales Percentage
MLS 53,625 81.44%


3,389 5.15%
FSBO/Pocket 8,830 13.41%

To place the percentage of historical sales listed on the MLS is difficult. We began matching all MLS sales to the recorded deed reporting that sale in 2012. Prior to 2012, we tracked public record sales and MLS sales separately. We believe the percentage of MLS listings hit their low point in 2005 when the supply of homes for sale hit their historical lows. We also believe the percentage of homes listed on the MLS peaked when bank and short sales dominated our market. Nearly 100% of all bank and short sales were listed on the MLS. My best guess is the low point for MLS listings was in the high 60 percentiles in 2005 and in the mid 90 percentiles when distressed sales dominated the market. The data below is for 2012 forward.

Value Proposition

A recent article published in RealtorMag looked at a study of price differences between homes sold by traditional agents and FSBO.

“The authors found that the differential in selling prices for FSBOs when compared to MLS sales is “remarkably close to average commission rates.” A FSBO sale, on average, nets nearly a 6 percent lower price than an MLS sale for a similar property.”

The report was based on a study done by Collateral Analytics. I have an issue with data scientists, as they don’t always understand the nuances of the data they are analyzing. In this case, based on the data they studied and the methodologies they used, I think they got it right. My only real issue is their simplistic definition of FSBOs and the message that conveys. In Arizona, we are fortunate to have our assessors provide us with full cash values (FCV), values held to strict equalization standards. By comparing the sale prices of resale homes to their FCV and averaging the individual calculations we can compare values. In laymen’s terms, properties sold on the MLS were 6.3% higher than properties sold off the MLS, confirming the conclusion of Collateral Analytics.

Thanks to the Information Market, we have additional data points not available to Collateral Analytics that allow us to view other variables. One of those variables is transaction type. Nearly 100% of all GSE (Government Sponsored Entity), bank sales and short sales take place on the MLS. By their very definition, they’re going to sell at lower values even though they’re listed on the MLS and the agent represents the best interests of their client. There were 450 short sales that sold at a factor of 1.1969, 380 GSEs that sold at a factor of 1.2736 and 539 bank sales that took place at a factor of 1.1729. Maybe the most interesting statistic centers around properties that were sold while being in foreclosure. In our system, we code these as Active Notice transactions. These are properties that are heavily solicited by listing agents and investors. When in foreclosure, you’re definitively better off selling your property on the MLS. Foreclosures listed on the MLS sold for 20.40% more than bargain stalking BOSAs and 10.10% more than FSBOs.

The ARMLS Pending Price Index (PPI)

Last month STAT projected a median sales price for August of $242,000. The actual median sales price was $245,000, missing the mark by 1.2%. Looking ahead to September, the ARMLS Pending Price Index anticipates the median sales price will fall in September, projecting a median sales price of $239,900. I expect the median price will be lower in September but not as much as our model suggests. It’s not unusual for the median sales price to fluctuate downwards in the fall months as seasonal patterns take hold.

Sales volume in August was 8,113, roughly 3.4% higher than the 2016 total of 7,843. Sales volume for the first 5 months of 2017 was 7.6% higher than 2016, with 65,147 sales in 2017 compared to 60,564 for the first 8 months of 2016. We begin September with 6,203 pending contracts, 3,581 UCB listings and 521 CCBS giving us a total of 10,305 residential listings practically under contract. This compares to 10,498 of the same type of listings one year ago. Even with fewer “pending” listings this year compared to last, I expect September sales to be higher than the volume of 2016. We reported 7,328 sales in August of 2016. Sales this September are expected to be in the 7500 range.

Written by Madison G from ARMLS
Madison has been part of the ARMLS family since 2015. As the Content Editor, she brings an extensive background in writing, editing, communication and punnery.