Three researchers believe a first-of-its-kind dataset has enabled them to definitively answer a long-debated question: why have buyer agent commission rates been so stable historically?

In metros across the country, decade after decade, pre-internet and post-internet, buyer agents almost always get 2-3% of the sale price, paid by the seller, the researchers note.

One of the researchers, Will Fried, had a hunch. The senior data scientist at appraisal tech startup True Footage previously worked at REX, a mostly defunct discount brokerage that charged seller fees as low as 2.5% and relied on salaried agents rather than commission-paid agents, who typically cost sellers 5-6%.

REX, which just lost a federal case against Zillow, met significant pushback from agents, based on recorded calls with agents that the company made public, like the excerpt below.

“It’s your prerogative to do whatever you want to do, but I can’t aid in the demise of my profession. I’ve been an agent for 15 years, so I can’t aid in that. So I’m not going to show properties that do that.”

Unnamed agent in recorded call with REX

Not showing clients properties because they have below-market commission rates is an example of “steering,” a practice that is often alleged about buyer agents and just as often denied by broker firms and associations.

To go beyond REX’s anecdotal evidence, Fried teamed up with Jordan Barry, a professor of law and taxation at the University of Southern California, and John Hatfield, a professor of finance at the University of Texas – two researchers who have previously published studies on other “collaborative” industries where fees prove uniform and persistent.

The researchers seized on a dataset that prior to 2020 was never available to the public or researchers: the buyer commission rates sellers offer when they list their homes for sale.

Multiple Listing Services were previously prohibited from disclosing these rates by a National Association of Realtors rule, but the rule was repealed as part of a proposed settlement with the U.S. Department of Justice and portals like Redfin began displaying the rates soon after.

The researchers scraped this data from Redfin’s website over several months and amassed what they believe to be “the first systematic, nationwide evidence that buyer agents do in fact steer clients away from properties that offer low buyer commissions.”

Their findings, published online as a draft on Oct. 17, allege sellers who offer low commissions get fewer page views on their listings, take longer to sell their homes and face higher odds of not selling their home at all than sellers who offer the market rate.

Market uniformity

The researchers’ data offers the first nationwide glimpse of the rates sellers offer when trying to sell their home, and it is a picture of market uniformity.

The average rates offered in the 34 metros studied vary by less than 1 percentage point – from 2.22% in Boston to 2.93% in Austin and Columbus.

According to the study, this uniformity persists “when there are many home sales and when there are few,” when there are more agents or fewer agents and “even as the internet makes it easier than ever for homebuyers and sellers to find each other.” In other words, the rates seem impervious to market forces, even as the internet has obliterated the historical rates of other intermediaries, like travel agents.

So how is the uniformity and permanence of buyer agents’ 2-3% rate possible in a free market?

The researchers argue it is possible because buyer agents, informed of the rates each listing offers thanks to MLSs, direct their clients’ attention away from those offering below-market-rate commissions and towards those offering the market rate. That theory, which is fiercely denied by many in the industry, is not new.

It has been investigated several times by various government agencies, the researchers note, and is currently the subject of a DOJ investigation and a class-action lawsuit. Just weeks after the draft published, a jury in Missouri found NAR, HomeServices of America and Keller Williams guilty of colluding to inflate or maintain high commission rates in a separate class-action lawsuit over buyer broker commissions, known as the Sitzer/Burnett suit. With treble damages, the defendants could pay up to $5.36 billion in Missouri alone.

Regulatory actions from 1950 to 1971 focused on mandated or recommended fee schedules explicitly stated in NAR’s or affiliates’ policies. Actions since NAR banned mandatory and recommended fee schedules in 1961 and 1971, respectively, have often focused on allegations of agents steering clients.

The study claims many of these investigations have been hampered by a lack of data, a problem the researchers believe they have overcome.

The data

If buyer agents truly steer their clients, the researchers theorized, then Redfin listings with lower-than-market-rate buyer agent commissions should get fewer page views than those with market-rate commissions. That would likely indicate agents aren’t sending their clients links to those properties, the hypothesis goes.

At this point, the reader’s mind may start to spin: what about people who just browse properties online with no intention of buying? What about the difference in desirability between neighborhoods? What about homebuyers who look at Zillow or Realtor.com instead of Redfin? What about the skill or experience of the brokers involved?

Through statistical controls and the exclusion of properties without commission data, within the top or bottom 5% of home values, without a completed structure, etc., the authors attempt to address those concerns. The full methodology is available in the draft.

After the exclusions and tinkering, the researchers were left with about 265,000 listings across 34 metros with data pertaining to price, building features, page views and more captured between June 2021 and February 2022.

Redfin, however, is skeptical of the data the researchers scraped and told HousingWire that the researchers did not tell the company they were using their data for the study.

“If they had informed us that their study relied on scraped page views data — which we removed from Redfin’s website more than a year ago — we would have informed them that the data was inaccurate,” Redfin Communications Director Angela Cherry said by email.

A Redfin webpage says page views were removed from the site on July 1, 2022.

“We stopped displaying page views data after we discovered that the system calculating them was not scaling with our growing website traffic, and therefore was unreliable,” Cherry continued. “We plan to resume publishing the data when an overhaul is complete and we’re confident in its accuracy.”

The researchers remain confident in the credibility of the page view data, they told HousingWire. They validated the data with a number of tests and point out that page views are higher in desirable neighborhoods and for homes that are underpriced, trail off the longer the listing is online and otherwise behave as one would expect.

Furthermore, the analysis relies not on the absolute page views each listing gets – which may change when Redfin rolls out its new methodology – but on the page views relative to other listings. Page views also are only one leg of the study’s three-legged stool; the researchers examined commission rates’ impacts on page views, days on market and sale probability.

The result of this analysis was clear, they said: Sellers who offer below-market buyer agent commissions get fewer page views on their listings, take longer to sell their homes and face a higher probability of never selling their home.

But what about the internet?

The reader still may have another objection buzzing in their mind: can agents really steer clients when the clients can simply search for properties online?

In previous statements, NAR has claimed sites like Zillow and Realtor.com have nullified any ability agents may have previously had to steer clients. Redfin, too, told HousingWire, ” Only a small share of Redfin’s traffic and listing views come from real estate agents forwarding listings to their clients. The majority comes from our customers’ own searches as well as automatically generated alerts that we send to customers.”

The study’s authors addressed this concern by conducting a survey. Of the 184 respondents who purchased a home within the last five years:

• 92% said their agent e-mailed or texted them properties of interest
• of that group, 90% reported viewing them on public portals such as Zillow, Trulia, Realtor.com and Redfin
• the median buyer viewed 70% of the listings their agents recommended

NAR’s own research reinforces the claim that agents have steering power, the researchers note. NAR’s 2023 Home Buyers and Sellers Generational Trends Report states that 29% of home buyers purchased a home their agent found, and “help find[ing] the right home to purchase” was the top result when buyers were asked what they want most from their agents.

Even when a client finds a property they want to see, opportunities for steering remain, the study notes. In one of REX’s recorded calls with a broker, for example, a broker who was frustrated with REX’s fee process says that they would simply tell their client – who found the REX property on Zillow – that the property had sold.