If you read Inman News, it is likely that you have seen the story on Redfin's blog entitled "Sellers Lose in Dual Agency." Glen Kelman, CEO of Redfin, had his firm look at the data covering 230,000 closed deals and concluded that homes sold in what REDFIN deems to be dual agent transactions sold for 1.6% less than transactions where a buyer agent was involved. In Redfin's post, Mr Kelman was gracious to link to my own counterpoint from a little while back. He disagrees, but does not address any specifics.
I don't know about you, but if I were a consumer and read that I could get a deal for 95.6% of asking through the listing agent instead of 97.2% with my own rep, I'd fire my buyer agent and just deal directly with listing agents. I don't think he meant to do this, but I don't see how many members of the public would not unwittingly draw this conclusion.
Part of my comment follows:
In one of those transactions, a Redfin buyer client, after months of looking with me, ended up wanting to see one of my own listings. On the advice of Redfin's Kevin Broveleit, I stepped back and designated an agent in my firm for both buyer and seller, as I am principal broker. This was an "in house sale" where I got MLS credit on both sides, but was not the type of dual agency Mr Kelman refers to. The devil in the details is that Redfin is assuming that MLS credit to one agent includes in house sales where buyer and seller had designated agents (which Redfin does engage in themselves) or simply a buyer customer who did not have the fiduciary service from the listing agent whose client remained the seller. Neither of those cases is the dual agency Redfin warns against.
Full disclosure: I was Redfin's first partner agent when they opened business in New York. I remain good friends with Michael Daly, who is, in my view, a valuable asset to their enterprise and a top shelf man.
In New York (which were three of the largest counties in the Redfin study), concluding that MLS credit to one agent in a transaction is dual agency is supremely erroneous. It could have been an "in house" sale where both buyer and seller had their own designated agent-something Redfin themselves engage in- or the buyer was simply a customer (as opposed to the seller client) who had no fiduciary representation from the listing agent. I am sure this is the case in many areas outside of New York. In order to know if there was really dual agency, you have to open the confidential file and read the agency disclosure. I've met Glen twice and we shook hands, but I never showed him any file, even on deals where Redfin referred me business.
The data that Redfin uses to come to their conclusions is their IDX or Virtual Office feed from the MLS systems they belong to. How do I know this? They told me. This data is meant for consumer searches. It populates websites where consumers look for homes. It isn't really a data pool that is meant for such a deep drilling matter as agency disclosure. A few months ago, when Redfin attempted to publish "field reports" of agent statistics, they had to end the project because the accuracy was a problem. I believe the same problem occurs in this instance when they are using a pool of data that is not being utilized for the its true purpose. There simply isn't enough detail in IDX/VOW data feeds to analyze it properly for something as sensitive as agency representation and the consequences of such.
For example, as I comented:
Do overpriced listings, which end up stale, sell for a higher percentage than aggressively priced homes that garner vast attention from cooperating firms? I think we know the answer. So, yes, listing agents often end up selling their overpriced inventory, (if they sell at all) themselves. And some of the "lowballers" do deal directly with the listing agent as part of their strategy. Even in those cases, with the seller as client and the buyer as customer, it is invalid to conclude that dual agency occurred.
The devil is in the details. Without facts beyond raw sales and MLS credit, you can't normalize the data for crucial variables, which is a basic statistical principle. Rather than paint with a broad brush over 230,000 deals, I think there would be more value to look carefully at 100 or 250 transactions. I don't believe for a minute that buyers get a better deal with the listing agent directly, and I don't think Glen would build a business modeled on that either.
Redfin would do better to build their brokerage to be profitable and sustainable so that they never need another dime of venture capital, and leave the statistical analysis to the people with superior data. Those people would probably not be brokers. Any analysis that makes a buyer conclude they'll get a batter deal without a true representative does those buyers no service.