NBER Study: Conflicts of Interest and the REALTOR Commission Puzzle

Yesterday, I gave a presentation at the Pacific West Association of REALTORS Broker Summit which I named “Bring the Pain”. To be fair, I ask all my speaking clients whether they want the Light, Medium, or Heavy and PWR picked heaviest of the heavy.

One of the things I mentioned was a recent Working Study published by the National Bureau of Economic Research (NBER) called “Conflicts of Interest and the REALTOR Commission Puzzle”. I’ve embedded the PDF after the jump, and I had promised the attendees that I would make the paper available. I figured I might as well make it available to all my readers, who are the best informed readers in the real estate industry. 🙂

Cut to the chase: this study is truly disturbing, if it is valid. Basically, it undermines the justification for buyer agency and lays the groundwork for regulation/legislation banning the sharing of commissions. So yeah, it’s worth paying some attention to….

The Working Study

First of all, here’s the full paper:

Read the whole thing, although to be fair, unless you’re an economist or a mathematician, huge tracts of the paper will be as if it were written in ancient Greek.

Second, the authors of the study are what we might call formally as “Really Smart People”:

Relatedly, NBER isn’t some brand new think tank of Social Justice Warriors. It is one of the most respected economic research think tanks in the United States founded in 1920. From their About Us:

The NBER is the nation’s leading nonprofit economic research organization. Twenty-four Nobel Prize winners in Economics and thirteen past chairs of the President’s Council of Economic Advisers have been researchers at the NBER. The more than 1,400 professors of economics and business now teaching at colleges and universities in North America who are NBER researchers are the leading scholars in their fields.

In fact, NBER is the organization that calls when a recession starts and ends. It is extremely influential in economic policy circles and with policymakers throughout government and business.

So, what did this study actually say?

The Findings

Assuming that most of you didn’t actually download and read the paper above, let me summarize their findings. First, the Abstract:

This paper documents uniformity in real estate commission rates across markets and time using a dataset on realtor commissions for 653,475 residential listings in eastern Massachusetts from 1998-2011. Newly established real estate brokerage offices charging low commissions grow more slowly than comparable entrants with higher commissions. Properties listed with lower commission rates experience less favorable transaction outcomes: they are 5% less likely to sell and take 12% longer to sell. These adverse outcomes reflect decreased willingness of buyers’ agents to intermediate low commission properties (steering) rather than heterogeneous seller preferences or reduced effort of listing agents. While all agents and offices prefer properties with high commissions, firms and agents with large market shares purchase a disproportionately small fraction of low commission properties. The negative outcomes for low commissions provide empirical support for regulatory concerns that steering reinforces the uniformity of commissions. [Emphasis added]

In plain English, what the above says is that properties whose cooperating compensation is below the norm for a market (e.g., if 3% is standard, then a 2.5% commission) are 5% less likely to sell at all, and even when they do sell, take 12% longer to sell. The only explanation for this — according to the authors — is that buyer agents are steering consumers away from lower-commission properties and towards higher-commission properties. Given that, they think their findings support “regulatory concerns.”

What might this regulatory concern be all about? Well, from the Conclusion:

Our findings provide empirical support for regulators’ long-standing concern of steering behavior contributing to the lack of variation in commission rates, despite consumers’ increased access to information and lower search costs due to the internet (GAO, 2005; FTC, 1983, 2007).

Compared to other industrialized countries, commission fees in the United States are high. For example, commission rates average less than 2% in the United Kingdom and the Netherlands, compared to the typical rates of 5% and 6% in the United States (Delcoure and Miller, 2002). As highlighted by our model, unbundling commissions has the potential to eliminate steering and reduce commission fees. Given the sheer size of aggregate housing transaction values, even modest reductions in commission fees could lead to a non-trivial reduction in transactions costs. [Emphasis added]

Again, translated into plain English mixed with industry jargon, what the above says basically is this:

Why are real estate commissions so high, despite low barriers to entry and all the advances in technology that should have reduced the cost of brokerage services? It’s because cooperation and compensation keeps commissions high through steering by buyer agents. The FTC and other regulators have been and are very concerned that consumers are getting ripped off, and this study shows that they are getting ripped off.

My oh my.

The Justification for Buyer Agency, Undermined

I’m not even going to bother trying to look at the meat of the study, which is pretty intensely quantitative. Here’s just a sample:

We refine the comparison in Figure 4 by controlling for firm attributes in the following regression:

1(TopRevlmt) = γ frcRtL25lm,t−1 +Xlm,t−1β +μmt +εlmt, (1)

where 1(TopRevlmt) is 1 if office l’s revenue is in the top quartile in market m and year t, X represents office controls and μ represents market-year fixed effects. The key regressor is frcRtL25lmt, the fraction of office l’s listings that is below 2.5 percent in the most recent three years t − 2 to t. The one-year lag of frcRtL25lmt alleviates concerns that it might be jointly determined with the dependent variable.

If you’re an actual economist, or a mathematician, please feel free to review the Study and let the rest of us know how valid or erroneous it might be. I know I couldn’t translate the above to anything.

Let me instead address the implications of the study, assuming that the authors’ findings are indeed valid and accurate for the 653,475 listings they analyzed.

Basically, this destroys the justification for Buyer Agency and undermines cooperation and compensation.

Buyer Agency was created out of sub agency because buyers were confused as to who was representing their interests. The key change was to make the buyer agent a fiduciary of the buyer with all of the heavy responsibility that entails.

The basic concept of a fiduciary is that the client’s interests come first and foremost in every situation. Well, if the Study’s findings are true and valid, then it appears that huge numbers of buyer agents are violating their fiduciary responsibilities. That’s what ‘steering’ means. The authors claim to have controlled for property quality and listing agent effort, so the 5% and 12% negative impact is explained by buyer agents steering their consumers away from the “low-commission” houses and towards the “high-commission” houses.

So that’s one.

More profoundly, the findings of the study suggest that cooperating compensation itself is harmful to consumers. The authors call that “bundling” of commissions. If that “bundling” results in steering effects that lead to artificially high commission rates, then well, we’ve got a problem. The Authors actually put numbers to that:

The calculation above suggests that sellers choosing the high commission rate pay $4455 more (roughly 1% of the sale price) in exchange for 20 fewer days on the market compared to sellers who initially list at a low commission rate.

Let me suggest that this is Not Good News.


Don’t Panic

At this point, there is no reason to panic. This is just a Working Paper. Yes, it’s by three very smart economists, and published by the leading economics think tank. Yes, NBER is very influential, but influence doesn’t mean direct power.

NAR still exists, and still remains very powerful, and it isn’t clear that anything will come of this Study. So relax. Don’t panic. At this point, this paper is interesting intellectually, and as a matter of tracking where the intellectual Elites are.

At the same time, it’s worth keeping an eye on, because the chance that average consumers and average voters would oppose any regulation/legislation that unbundles real estate commissions is nil. And there are people like Richard Cordray of CFPB (who is completely unaccountable to anybody) who could simply ban cooperation and compensation by fiat.


Share & Print

Picture of Rob Hahn

Rob Hahn

Managing Partner of 7DS Associates, and the grand poobah of this here blog. Once called "a revolutionary in a really nice suit", people often wonder what I do for a living because I have the temerity to not talk about my clients and my work for clients. Suffice to say that I do strategy work for some of the largest organizations and companies in real estate, as well as some of the smallest startups and agent teams, but usually only on projects that interest me with big implications for reforming this wonderful, crazy, lovable yet frustrating real estate industry of ours.

Get NotoriousROB in your Inbox

22 thoughts on “NBER Study: Conflicts of Interest and the REALTOR Commission Puzzle”

  1. Yes, lower commission rates do equate to longer marketing times on occasion however I sincerely beg to differ to these think tank genious’! Lower commission rates in my world generally indicate lower/poorer service to that particular seller that equates to homes NOT READY for market! Poor photos, bad curb appeal, improper/no staging and the list goes on! In order for these think tank guru’s to make any qualified opinion regarding real estate they have to get out of the building, look taste and feel real estate! Then tell me WE make to much money!!

  2. RSH~ enlightening to say the least, and very informative; thank you! The problem is this issue is not so in your face and it is very hard to prove, and that’s a shame! But shame on all and any acting out agents that refuse to show listings because of a lower commission! Bottom line “pigs get slaughtered” and (IMO) for any pig agent acting in such a manner they don’t deserve to hold a license.

    BTW: I am in commercial real estate, and I am always looking to buy apartment buildings. And there is a similar problem going on with the exception listing agents are to blame, and there’s no deception “listing agents are bold about it.” Companies such as Marcus & Millichap double end their deals. Meaning, if I want to buy an apartment building listed by M & M, I can’t represent myself. Even “if” I am willing to write up a solid full price offer, the listing agent won’t allow it or present it! The listing agent tells me “if” I want to buy the building, they have to represent me on the deal.

    I would love to hear your thoughts on this type of agency behavior! As you might imagine “I can’t stand it and on so many levels I feel it is so wrong.”

    • Shenanigans in commercial real estate is as old as the Code of Ethics is in the residential side. 🙂 I’m not in the least bit surprised, but then, CRE is a totally different animal from RRE.

  3. It appears they have overlooked (and so failed to control for) the propensity for some sellers to seek out discount and low commission brokers while at the same time insisting on over pricing their homes.

    • Actually, they write this:

      We further probe the robustness of these results by adding measures of listing office and agent quality. These additional controls alleviate concerns that lower quality offices or agents are more likely to list at low commission rates (columns 5 and 6). For agents, we control for their experiences over time and also whether they are star agents (ranked in the top decile using agents’ average annual listings). For offices, we control for the composition of agents in the office, the performances of listings by the office in each year (such as the fraction of listings that were sold, the average days on market for sold listings) and whether an office is the dominant office in a market in terms of average transaction volume. Higher quality offices and agents have higher sales probabilities through two channels. First, they are better at selecting properties that are easier to sell. Second, they are more knowledgable about local market conditions, have better social skills, and are better at selling.

      • Seriously? They are controlling for all that? If it is true my mind is totally boggled. We cannot control for agent quality but they are “teasing it out” and somehow understand how to pull that out of the equation?

  4. In addition to what others wrote, did the paper take into considerations that PERHAPS the higher rate listings were more highly advertised and marketed? Sherry above noted that maybe the lower listings weren’t as good but maybe it’s the opposite that the higher paying listings offered more. Maybe more than just being out into the MLS with a sign in the yard? Maybe the higher listings are by agents that work that market more often and know the other agents who are likely to have buyers for their listings and sell them. Broker tours? Like others said, there are too many variables to just “account for” in a paper. With 20 years in this business I believe its a very small percentage of agents who look at commission rates first (if it’s even 5%, I’d be shocked).

    • Jim, I’m far from qualified to make sense of some of the mathematical formulas and the regression analysis that these three econ professors conducted. But they claim to have adjusted for “listing agent effort” in their formulas as well as “quality of listing”. Have no idea how they did that. Like I said, if someone has the background to decipher what the authors wrote, I’d love to hear a counter/critique of their findings.

  5. This may be a case of correlation does not imply causation….and/or “Lies, damned lies and statistics.”

    Sherry and Paul bring up valid points. Is there a way to account for the original listing price vs. sales price? Those with the larger price gap could then be compared to the DOMs to similarly over-priced homes listed with the “standard” commission.

    It’s hard to imagine: “We saw this house and Zillow and we’d like a showing, please.” “Um, no, that property has a lower commission…is on a fault line.”

  6. The economists came in with a preconceived idea, and set out to prove it. Sellers who offer a lower commission are the same who price too high and don’t fix up their house in advance – all symptoms of low motivation. Those are the listings that take longer to sell, or don’t sell at all.

  7. Sherry, Paul, Jim, Robert — like I said, I’m not qualified to judge the work of these economists, but they do write this in the full report:

    Lower sales probabilities for low commission listings might be driven by seller preferences. In particular, we are concerned that patient sellers who are more likely to trade off higher sales prices against lower sales probabilities are also more likely to list at low commission rates (to maximize their proceeds net of commission). In column 4, we proxy for seller patience using the idea that patient sellers will list their properties at higher prices, relative to prices predicted from observed attributes. This also builds on the notion of using the list price as a proxy for the seller’s reservation price (Genesove and Mayer, 2001). Patient sellers tend to have higher reservation prices than motivated sellers. We first calculate the ratio of the observed list price to a predicted hedonic price, then construct decile dummies for this ratio. These decile dummies constitute our seller patience controls. With these controls, the effect becomes slightly less negative (- 6 p.p.). We explore other controls for seller patience and seller preferences in Table 7 (see Section 5.2.2 for details) and reach similar conclusions.

    Make of that what you will.

    • Had they used “hedonistic” vs “hedonic” value in their equation, they would have been more on track. 🙂 LOVE this study! I will be reading every word of it!

      Thanks for the heads up!

  8. The assumption that buyer agents are steering their clients away from properties with a lower offer of compensation is a leap that should not automatically be made. If a brokerage has a buyer representation agreement with a buyer we would (in most states) have to disclose to the buyer prior to showing the property that this is one where the buyer would owe us some part of the compensation. At that point the BUYER often chooses not to see the property. To assume that the agent just didn’t show it is, as I said, a leap. I am not naïve enough to believe that this type of steering sometimes happens, I think assuming it happens all the time is jumping to erroneous conclusions.

    • Good point about the buyer agent contract but they wouldn’t have had access to that type of data to using in their independent variables.

      I do think, though, that there is something to be said for decoupling the buyer’s agent commission from the listing.
      I’m an exclusive buyers agent and I would welcome it – so long as the states I’m licensed in required ALL agents representing either side to have a contract specifying the commission they would be paid – as a condition for that payment. They should also require that buyers provide some minimal amount of identification before they may enter a seller’s home.

  9. I don’t see that this really controls for seller motivations or ‘patience’.
    For anyone that wants to look at the article you can down load it (for now) from here:


    Page 39
    Table 7: Robustness check, controlling for seller preferences
    Dependent variable: Probability of being sold
    (1) (2) (3) (4)
    Specification List Finer Seller No common
    price patience controls name names
    Low commission listings -0.06*** -0.05*** -0.07*** -0.07***
    (0.004) (0.003) (0.02) (0.02)
    N 344832 344832 31407 30120
    R-squared 0.50 0.51 0.48 0.49
    Market-year FE, month FE
    Property, agent, office controls Y Y Y Y
    Property FE Y Y N N
    Ln(List price) Y N N N
    Seller patience N Y Y Y
    Seller FE N N Y Y
    * p<0.1, ** p<0.05, *** p<0.01
    Notes: OLS regressions at the listing level for the effect of low commission rate on the probability of sale, with different
    specifications to control for seller preferences. Columns 1 and 2 are similar to the most saturated OLS specification in Table 3
    (column 6). Column 1 controls for ln(List price) instead of seller patience deciles. Column 2 controls for seller patience using
    one hundred percentile dummies. Column 3 includes 14213 seller fixed effects (defined using seller names). This specification
    restricts the sample to sellers with multiple listings and seller names that could be identified using the county records. Column 4
    is similar to column 3, but drops common names (names that occurr more than 5 times in our data).

  10. The study’s conclusion of steering is based on “less favorable transaction outcomes” But, there are other elements of the transaction that precede the “outcome”. No mention was made of comparing the number of showings of “low co-op” vs “conventional co-op”. As mentioned above, maybe for a number of reasons the “low co-op” listing was not as attractive as other listings.

    • I didn’t see seller concessions as a result of home inspection negotiations shown there as a factor either. Of course, they would not have access to that or be able to separate it out of total seller closing cost concessions.

  11. Often lower commission agreements are made by either
    A) Offices that don’t put their homes on MLS OR
    B) Limited service offices where the seller pays a fee to go into MLS but all showings, negotiation, etc are done directly with a seller who has overpriced their home.

    I wonder if economists use one side of their brain to make an analysis of the process of a home purchase and home buyers (and agents) are using the opposite. They want a house in a good location, that “feels” like home in their price range and that’s what we show them.
    We, as agents, can’t steer a buyer one way or another if the home fits their needs, they are seeing it all on the internet practically before we are!

  12. As you have stated many times, the internet has changed everything about home selling. And 1998-2011 is as current as they could get? The dark ages. How did they control for Avocado green appliances?? Or dirty diapers ? Or tenants that won”t let you into the home……or snow??? Case Schiller at least uses the same metric for the same house. Typical egghead production, “we controlled for an over priced listing” or “we controlled for the unrealistic seller” Sure you did.

Comments are closed.

The Future of Brokerage Paper

Fill out the form below to download the document