Real Estate Brokerage Financial Reports

Document Sample
Real Estate Brokerage Financial Reports Powered By Docstoc
                         JOHN C. WEICHER*

   This paper discusses the available information on the cost of real
estate brokerage to consumers. Economists, at least, are particularly
interested in the question of whether the brokerage market is com-
petitive, a question that has become especially important in the
context of the Department of Justice anti-trust suit against the
National Association of Realtors (NAR), concerning internet access
to Multiple Listing Service (MLS) databases.1
   There is a fairly widespread view that brokerage is not a compet-
itive industry, a view shared by publications as diverse as the Wall
Street Journal and the New Republic.2 This opinion is based on sev-
eral perceptions: commission rates are ‘‘too high’’ for the work
involved, and are sticky downward even as technology reduces the

    John C. Weicher is Director of Hudson Institute’s Center for Housing and
Financial Markets. Dr. Weicher would like to thank Xiuyue Zhu for research
assistance; Jesse Gurman, Norman Hawker, Austin Kelly, and Thomas G.
Thibodeau for providing copies of various studies; and Peter F. Colwell for
helpful comments on the literature.
    U.S. Department of Justice (2005).
    See for example Wall Street Journal (2005), Murray (2005), Hagerty
(2005), and Risen (2005). Evans (2005) lists several other publications, includ-
ing USA Today, Computer World, the Washington Post, Time Magazine, and
Forbes, and also ‘‘The Tonight Show’’ and ‘‘The Today Show’’ on NBC.


120                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

broker’s cost;3 commission rates are higher in the U.S. than other
countries; NAR and state Realtor associations lobby vigorously for
state laws to restrict competition, for example by setting minimum
commission rates and by requiring that brokers provide ‘‘full ser-
vice;’’ NAR has successfully lobbied Congress to prohibit banks
from entering the real estate brokerage business; and NAR restricts
access to the MLS by discount and internet brokers.4 In response,
industry advocates argue that: commission rates are highly variable
and have been declining for over a decade; the number of real estate
brokers and agents is increasing, and entry to the industry is easy;
brokerage practices and state regulations are hardly limiting the
market, since existing home sales are at record volumes;5 real estate
brokers bear substantial marketing costs with no guarantee of a sale;
bank entry into brokerage will beneŽt the banks and the Federal
treasury, but not consumers; and homes are unique while other
goods sold over the internet, such as stocks and airline tickets, are
‘‘a commodity where one share or ticket is the same as the next.’’6
   A number of these issues are being addressed by other papers at
this conference. Hahn and Litan describe a number of state laws and
regulations and industry practices that appear too many economists
and antitrust lawyers to be anticompetitive trade restraints, and have
analyzed their impact.7 Delcoure and Miller compare brokerage
costs in the United States with other developed countries.8 Lande
and Marvel address an important question that must occur to any
economist (and has been studied by several): how a cartel can be ef-
fective with free entry into the industry.9 This paper therefore
excludes these subjects, although the review of the literature neces-
sarily impinges on several of them.
   My focus is empirical. A number of economists have created
formal models of brokerage to illuminate various aspects of the
brokerage market. Some have concluded that brokerage commis-

      Murray (2005); Bahney (2005).
      Hagerty (2005).
      Lereah (2005).
      Evans (2005).
      Hahn and Litan (2005).
      Delcoure and Miller (2000).
      Lande and Marvel (2000).


sions are well above the competitive market equilibrium.10 Others
have argued that there is ‘‘a simple economic rationale for the
observation that real estate commission rates seem to exhibit
remarkable rigidity,’’ in that higher commissions may result in
greater eort by the broker or agent to sell the house.11 To my knowl-
edge, neither of these models, nor any of a variety of others, has
been empirically tested; nor am I in a position to test them in this
paper. It is my belief that a model should be tested empirically, even
if the formal analysis yields insights that correspond to market
behavior in a qualitative sense, but that has simply not been possible
in most cases.
   The paper is essentially a review of the limited empirical
evidence. It Žrst reports the available time-series data on brokerage
commissions, with some attempt to interpret the data despite its
shortcomings. It then discusses the cross-sectional, primarily aca-
demic literature on commissions in various markets. There is not
much. The data are usually proprietary and not readily available to
the public or to academic analysts. This is undoubtedly the main
reason for the overwhelming preponderance of theoretical studies.
To quote one economist who attempted to obtain commission rate
data in order to test his model, ‘‘It’s been one stonewall after

        Time-Series Data: The REAL Trends Survey
   Apparently the only source of information over time on real estate
commissions is a survey conducted annually by the industry publi-
cation and communications company, REAL Trends. Unfortunately,
this survey is not publicly available. The data are compiled from the
larger real estate brokerage Žrms, which consider the information
proprietary and will not allow it to be published. REAL Trends has
historically made some information public, but changed its policy in
2004, according to Steve Murray, co-founder and co-editor, and
will no longer disclose it publicly: ‘‘Based on new privacy policies
we put in place last year for our entire range of publications we can-
not use the commission information publicly any further. So al-
though we still collect the data for all of these Žrms we cannot and
will not use it in any other fashion than for verifying the inputs from

       See for example Anglin and Arnott (1999).
       Bruce and Santore (2004, p.10).
       Roberts and Mara, quoting Bruce (2005).

122                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

brokerage Žrms.’’13 Commission rates are available only from sec-
ondary sources, which report average commission rates for one or
two years and identify REAL Trends as the source of the
information. I have compiled these reports primarily from the inter-
net; the data that I have been able to locate appear in Table 1, with
the secondary source identiŽed.
   Table 1 indicates that there has been a downward trend in the
average commission rate between 1991 and 2004.14 Roughly speak-
ing, the average commission was about six percent or a little more
in 1991, and also about six percent in 1995; it declined to about 5.5
percent by 2000; then it dropped sharply to slightly above Žve
percent in 2001 and has remained at that level since then.
   There are several limitations to these data, and some reasons to
question the accuracy of the information as reported. REAL Trends
surveys the largest brokers, the ‘‘Top 500,’’ including ties. (The
actual number of Žrms in the Top 500 was 532 in 1998 and 541 in
2003, according to GAO.) These Top 500 Žrms accounted for 27.6
percent of all new and resale home transactions in 2001, according
to REAL Trends. It is not necessarily true that the largest Žrms are
typical of the industry.
   Besides the Top 500, REAL Trends collects information from an
additional group of ‘‘Up-and-Comers,’’ ranking in size just below
the Top 500. The number of Up-and-Comers seems to be variable;
it was 335 in 2003.
   The average home sales price reported by REAL Trends is high
relative to the market. In both years, the REAL Trends average was
about 13 percent above the average for existing homes sold, as
reported by NAR, and about 0.5 percent above the average for new
homes sold, reported by the Census Bureau. The REAL Trends
Žgure was about 11 percent above the weighted average for new
and existing homes combined in 2002, and about 9 percent in 2003.
(See Table 2.) A possible partial explanation is that the NAR series
is limited to single-family homes, while REAL Trends apparently
includes condominiums, but this is hardly enough to account for the
dierence. It appears that the REAL Trends survey is weighted to-
ward more expensive homes as well as large brokers.
   Besides the inherent limitations in the survey itself, there are
some discrepancies in the reported data. Although the survey results

       Roberts and Mara (2005).
    The survey apparently began in 1988, but the Žrst year for which I have
found any data is 1991.


are proprietary and REAL Trends does not make them public, there
is nonetheless a web page posted by REAL Trends reporting some
summary information for 2002 and 2003, apparently taken from the
May 2004 issue of the REAL Trends monthly newsletter.15 These
data include average commission rates by Žrm size and by region.
They are shown in Table 3. Although various published secondary
sources report an average commission of about 5.1 percent for both
years, the actual average commission is below 5.1 percent only for
the top 100. It is well above 5.1 percent for the remaining 400 Žrms
in the Top 500, and also for the Up-and-Comers. The average com-
mission rounds to 5.1 percent only if the top 100 and the remaining
400 have about equal shares, or the top 100 accounts for more than
half of the transactions for the entire group.16 Some further reason to
question the 5.1 percent also appears in the regional data. In both
years, the average commission is below 5.1 percent only for the
Middle Atlantic and Far West regions, which together comprise 11
states and the District of Columbia. Southern and Midwestern states,
which account for about 60 percent of annual existing home sales,
have average commission rates well above 5.1%. If the regional
Real Trends commissions are applied to the NAR regional distribu-
tion of existing home sales, the 2002 average commission would be
between 5.27 and 5.40 percent, and the 2003 average commission
would be between 5.19 and 5.31 percent.17
   A further complication arises because the same average commis-
sion rate is cited by dierent sources for dierent years. For

   These       data      are    at     ( –
     The varying data on coverage by size of Žrm may cause some misinterpre-
tations of the average commission. The veteran real estate reporter Ken Har-
ney, for example, reports that the average commission in 2003 was 5.06
percent (Harney, 2004), based on data from ‘‘nearly 900 of the largest real
estate brokerage Žrms across the country.’’ This is exactly the reported aver-
age rate for the Top 50 Žrms, and is lower than the average for any larger
number of Žrms. Harney accurately reports the average commission rates by
region that appear on the website.
      These calculations match the REAL Trends and NAR Midwest regions
and match the REAL Trends Southeast region with the NAR South region.
The lower numbers multiply the REAL Trends Middle Atlantic average com-
mission by the NAR Northeast region share of existing home sales, and the
REAL Trends Far West average commission by the NAR West region share;
the higher numbers multiply the REAL Trends Northeast average commission
by the NAR Northeast region share of existing home sales, and the REAL
Trends Southwest/Mountain average commission by the NAR West region
share. The regional divisions are not identical (Texas and Virginia are in dif-

124                       REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

example, Lehmann (2005) states that the average commission in
2000 was 5.48 percent, but Kannon Consulting (2004) assigns this
average to the year 2001, when other sources state that the average
was about 5.1 percent. This confusion may occur because REAL
Trends itself refers to surveys both by the year for which the data
are collected and the year the survey is reported. The data reported
in Table 3, for example, are listed by REAL Trends as ‘‘the 2004
REAL Trends 500,’’ when published in April 2004.18 For this rea-
son, when the same Žgure is cited by two dierent sources for two
dierent but consecutive years, I have assigned it to the earlier year
in Table 1.

                       Interpreting the Data
   During the last 14 years, home prices have risen while commis-
sion rates have apparently been falling. The nominal increase in the
average existing home price has outpaced the decline in the average
commission reported by REAL Trends, so the dollar value of the
average commission has been increasing. However, when house
prices are adjusted for ination, the real dollar value of the average
commission declined from 1991 through 1998, and then rose in most
years until 2004. These data are shown in Table 4. While the aver-
age commission rate declined by 16 percent between 1991 and
2004, the average commission in real dollars increased by 11
percent. The average real commission declined by about seven
percent until 1998, and then rose by 19 percent through 2004.
However, it was not until 2003 that the average real commission
exceeded the 1991 level.
   During the same period, there were large, steady increases in
existing home sales volume as well as existing home prices. Be-
tween 1991 and 2004, nominal average prices nearly doubled, and
the total dollar value of existing home sales nearly quadrupled. The
16 percent decline in the average commission reported by REAL
Trends, and the 39 percent ination over the period, together oset
part of this increase; total nominal commission income may have
risen by 230 percent, and total real commission income by about
140 percent. NAR reports a substantial but much smaller increase in
the annual nominal incomes of both brokers and agents; median

ferent regions in the two groupings), but these dierences do not aect the ba-
sic conclusion.
      However, Kannon also reports that the 2002 average commission was
5.12 percent, which is consistent with Lehmann and also Roberts and Mara.


agent income more than doubled, and median broker income
increased by about 75 percent, from 1992 through 2002.19
    Thus it might be reasonable to expect a substantial increase in the
number of agents and brokers, especially since it is generally
believed that the occupational barriers to entry are low for real estate
agents and for brokerage Žrms.20
    The reported downtrend in commission rates has been attributed
variously to rising employment and to changing technology. Given
the quantitative and qualitative limitations on the commission data,
it is not worth attempting any serious discussion of either hypothesis.
But even if the commission data are taken at face value, there are
limitations in the employment data. The most consistent time-series
on employment come from two surveys conducted by the U.S.
Bureau of Labor Statistics (BLS), the Current Population Survey
and the Current Employment Survey. The former surveys house-
holds and asks about the occupations of members; it publishes the
total number of individuals in ‘‘real estate sales occupations,’’
including the self-employed as well as employees. The latter surveys
business establishments and includes employed workers only; it
publishes the number of workers in ‘‘oces of real estate agents
and brokers,’’ and also the number who are ‘‘production workers,’’
which may correspond to actual brokers and agents as opposed to
administrative sta.21 As might be expected, the total including the
self-employed is about three to four times the total of employees
only; other data suggest that the self-employed far outnumber
    The three series move very closely from year to year; the simple
correlation coecients are all above y.90. All three also show a
generally rising trend in employment since 1990, during periods

       Roberts and Mara (2005).
    GAO (2005, p.8).
     The CPS series is reported annually at
special.requests/lf/aa200x/aat11.txt for years since 2000, and
pub/special.requests/lfaa9x/aat11.txt for earlier years; years before 1995 are
not available at the BLS website, but have been provided by Sharon Cohany
of BLS. The COS series IDs are CEU5553120001 for total employees and
CEU5553120003 for production workers, available on the BLS website at
   Another BLS data series, the National Employment Matrix, reports self-
employed brokers and agents as well as employees, between 1983 and 1996.
The number of self-employed ranges from two to four times as many as the
number of employees.

126                      REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

when reported commission rates were stable and during years of
decline. Further, all show accelerating increases over time. The
increase was smallest between 1990 and 1995, when according to
the REAL Trends data commission rates were stable; it was greater
between 1995 and 2000, when reported commission rates were fall-
ing; and it was still greater between 2000 and 2004, when reported
rates stabilized again. It is hard to see these patterns as evidence that
rising employment has driven down commissions. If that were the
case, commission rates should have fallen further after 2000. Much
better information, particularly about commissions, will be needed
to test any hypothesis about employment and commission rates.
   With respect to technology, the data are not extensive enough to
test any relationship between technological changes and commis-
sion rates.23

               The Cross-Sectional Literature
   There is somewhat more empirical evidence on brokerage com-
mission rates from academic studies, but still one is struck by the
paucity of research. In their exhaustive bibliography Žve years ago
Benjamin, Jud and Sirmans (cited hereafter as the BJS bibliogra-
phy) list 14 studies of commission rates over a period of 18 years.
Nine of these are classiŽed as ‘‘theoretical papers with simulation
results,’’ one is a theoretical paper without further qualiŽcation, and
only four actually have empirical results.24 For all of the other topics
surveyed in the bibliography, empirical studies predominate. To my
knowledge, there are no later studies of commission rates, and the
only other earlier one is a report by the Federal Trade Commission
(1983).25 The data in this handful of empirical studies are for transac-
tions that occurred between 1975 and 1992; only one has transac-
tions later than 1987. GAO notes, ‘‘Our review cites a number of

     There is some reason to think that the internet’s impact on brokerage is
not necessarily negative, from the broker’s standpoint. Zumpano, Johnson and
Anderson (2003) Žnd empirically that homebuyers who search the internet are
more likely to employ brokers than buyers who do not. They do not investigate
the relationship between internet usage and commission rates. They speculate
that the internet may increase agent productivity, as well as changing the
agent’s role.
     Another empirical study, by Larsen and Park (1989), is included in the
references of the BJS bibliography (BJS, 2000:1, p. 275), but not listed or
discussed in the text.
     Neither GAO (2005) nor Hahn, Litan and Gurman (2005) mention
anything since the BJS bibliography; nor is anything more recent listed in


academic studies that date back many years because, in large part,
there is not a large body of more recent research on the real estate
brokerage industry.’’26 GAO goes on to say that the older and more
recent research Žndings are consistent with each other, and with the
statements in interviews conducted in the course of their study.
   The most extensive of these studies is also the earliest. The FTC
compiled information from HUD-1 forms for a number of metro-
politan areas in 1975, 1978, and 1979. Data from 16 areas was col-
lected in one or more years, for a total of 23 city/year samples.27
FTC reported the distribution of the commission rate data in each
sample;28 subsequently, Carney (1982) conducted some statistical
analyses of the data.
   FTC distinguished between ‘‘stated’’ commissions, those listed
as a percentage on the HUD-1, and ‘‘calculated’’ commissions,
where the HUD-1 reports the dollar amount of the commission as
well as the sales price of the house. Calculated commissions tended
to be lower, and showed more variability. FTC also conducted a
survey of a national sample of home sellers in 1979-1980, Žnding
that 85 percent were quoted a commission rate of either 6 percent or
7 percent. About nine percent of sellers ended up paying somewhat
less than the quoted rate when the transaction actually occurred.
   Table 5 summarizes the FTC data for the calculated commission
rates in individual markets. FTC reported the full distribution only
for sales of existing homes. The distribution shows the modal com-
mission rate, and the proportion of transactions that took place at
that commission. As the table shows, the mode was always an inte-
ger, either 6 or 7 percent. In all but one sample, more than half the
transactions took place at the mode; in eight, the proportion was 80
percent or more. On average across the 23 samples, 64 percent of
transactions occurred at the local mode. FTC did not report the usual
descriptive statistics for the distribution; those are taken from Car-
ney’s article.29 Not surprisingly, the mean rate is nearly always quite
close to the mode, and in all but three samples, slightly lower. The
standard deviations range from 0.4 percent to 1.5 percent; the un-
weighted mean standard deviation is 0.96 percent. There is some-

       GAO (2005, p.9).
     The FTC also collected a large sample of HUD-1 data on the state level;
because these cannot be attributed to individual markets, they are excluded
from this paper.
     FTC (1983, pp.44-52).
     Carney (1982).

128                       REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

what more dispersion among the smaller samples (shown at the bot-
tom of the table); for samples of at least 50 transactions, the
unweighted mean standard deviation is 0.90 percent. Eighteen of
the 23 samples have standard deviations between 0.8 and 1.2
   The 1975 FTC are drawn from all lenders located in the central
counties of each Standard Metropolitan Statistical Area; it is not
clear if the loans are also limited to homes in these counties. The
1978 and 1979 samples are limited to lenders located in the cities,
but the data extend beyond the city boundaries because lenders made
some loans outside them. In addition, the FTC separately sampled
listings from more than one local MLS in some markets. In Los An-
geles and Seattle, listed rates were highly concentrated at the same
rate for all MLS, but in Boston and the Twin Cities there were
dierences. About 70 percent of listings for Greater Boston carried
a rate of 6 percent, while listings for Quincy/South Shore were split
about evenly between 5 and 6 percent. Similarly, in Minneapolis, 85
percent of listings had a commission of 7 percent, while in St. Paul
listings were split about evenly between 6 and 7 percent.30
   Apart from the FTC survey, the limited academic literature can
be easily characterized. Professors of real estate and related Želds
have collected transaction data in the areas where they live and
work. The well-known aphorism, ‘‘all real estate markets are lo-
cal,’’ can perhaps be supplemented by, ‘‘all research on real estate
brokerage commissions is local.’’ These are studies of local markets
by local faculty or graduate students. A corollary is that all of the
studies necessarily pertain to college towns. (This also applies to
empirical studies of other real estate brokerage topics, such as the
commission split between the listing and selling brokers, and the
time a home is on the market before it is sold, as can be seen in the
BJS bibliography.)
   The descriptive statistics for the academic studies are summa-
rized in Table 6. Larsen and Park (1989) collected information on
669 home listings in Lincoln during the Žrst nine months of 1986.31
Of these listings, 433 were sold and the data include the actual sales
commission; the data in Table 6 refer to the actual sales only. Larsen
and Park report the selling broker’s share of the commission, not the

       FTC (1983, p.53)
    Both Larsen and Park were on the faculty of Wright State University in
Dayton when the paper was published, but Larsen had received his Ph.D. at
the University of Nebraska.


total commission. They report that the selling broker’s share is
nearly always 40 percent, with rare exceptions.32 Their data bear this
out. Larsen and Park, unlike the other studies, report the full distri-
bution of brokerage commissions in their sample. Some 84 percent
of sales transactions have a 2.8 percent selling broker’s share, imply-
ing a 7 percent commission.33 For purposes of comparison, I have
inated the data to represent the full commission in Table 6.
   At about the same time, Goolsby and Childs (1989) analyzed
commission rates in Knoxville, using two samples totaling 275
transactions from 1983 and 1987.34 They also looked at the share of
the commission to the selling broker, which in Knoxville as in
Lincoln was nearly always 40 percent. They do not report the distri-
bution of commission rates or the mode, but it appears that the stan-
dard commission in Knoxville was apparently six percent rather
than the seven percent in Lincoln. Goolsby and Childs report a mean
selling broker’s share of 2.36 percent, with a standard deviation of
0.33 percent, equivalent to a mean commission of 5.90 percent with
a standard deviation of 0.8 percent.
   A few years later, Sirmans, Turnbull and Benjamin (1991)
analyzed commission rates in Baton Rouge over 1985-1987, using a
sample of 1,275 transactions limited to two subdivisions.35 They
also do not report a mode or a distribution of commissions, but the
mean rate of 5.81 percent suggests a mode of 6 percent. The stan-
dard deviation is reported as 0.6 percent.
   The only other academic empirical study, by Sirmans and Turn-
bull (1997) for Baton Rouge between 1985 and 1992, does not report
descriptive statistics.
   Comparison of Tables 5 and 6 shows that the standard deviations
for the data in the individual market studies are generally lower than
for the FTC data. Only two FTC samples have standard deviations
as low as or lower than those reported by Larsen and Park and by
Sirmans, Turnbull and Benjamin. Only three are lower than the 0.8
percent reported by Goolsby and Childs; Žve others are also 0.8.

       Larsen and Park at 428 (1989)
     The same is true for 96 percent of those not sold (the dierence suggesting
that there may be some exibility in commission rates when there is a distinct
possibility of a sale).
      Goolsby was a professor of Žnance at the University of Tennessee at the
      Sirmans and Turnbull were professors of Žnance and economics, respec-
tively, at Louisiana State University, and Benjamin was a graduate student.

130                      REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

   All of the academic studies, including Carney, go beyond the
FTC survey methodology and attempt to identify factors that aect
commission rates. Carney analyzes commission rates by each factor
separately, while the other studies report results for multivariate
regressions with the rate as the dependent variable. The most
important results of these statistical and econometric analyses are
shown in Table 7.
   New homes are often marketed through dierent arrangements
than the standard brokerage model for existing homes. Builders
may sell the homes directly, through sales people who are employ-
ees, or by contract with an independent brokerage Žrm which
handles all homes in a subdivision.36 Builders may therefore be in a
position to pay a lower commission because of the volume of busi-
ness they are oering the Žrm, and to simply pay their employees
less than an independent broker would charge. All of the studies
except perhaps Sirmans, Turnbull and Benjamin include new homes
as well as existing homes.37 (The FTC data in Table 5 is limited to
existing homes, but the FTC also collected data on new home sales,
which Carney includes.) All Žnd that commissions are higher for
sales of existing homes. In the 10 FTC samples where the sample
size for both new and existing homes is at least 20 transactions, the
unweighted mean dierence is 1.6 percent. In Baton Rouge, Sir-
mans and Turnbull Žnd a dierence of about 0.6 percent. Goolsby
and Childs treat age of home as a continuous variable, so a similar
calculation is not feasible. Carney also Žnds more dispersion in com-
missions for new homes; the standard deviations are higher by 0.7
percent, on average, almost double the mean standard deviation for
existing homes, and the dierence is signiŽcant in seven of the 10
   A second factor is whether the same broker both lists and sells
the home. Listing brokers typically oer a share of the commission
to the broker who brings in the buyer (60 percent in Lincoln and
Knoxville in the 1980s, for example). For some sales, the listing
broker is also the selling broker, and the commission may be lower;
Carney suggests that listing brokers may withhold listings from the
MLS if they expect to be able to sell the home relatively quickly,

    Carney (1982).
    Sirmans, Turnbull and Benjamin do not state speciŽcally that their data
are limited to existing homes, but all transactions occurred in two subdivi-
sions, and the mean age of homes sold was seven years. Whether the data
include new homes or not, the analysis does not include age of home as an in-
dependent variable in the regression explaining commission rates.


thus incurring lower costs; they will also be able to keep the full
commission. Carney terms the case where the listing broker works
with a selling broker as a ‘‘co-op’’ transaction, and the case where
the listing broker is the selling broker as a ‘‘non-co-op’’ sale.
   Carney Žnds that co-op sales have higher commission rates in all
of the FTC samples, and signiŽcantly higher rates in eight of the 16
where the sample size in each category is at least 20 transactions.
The unweighted mean dierence is 0.9 percent for the 16 larger
samples. Goolsby and Childs, however, Žnd no signiŽcant dier-
ence, to their surprise. The two Baton Rouge studies do not
distinguish co-op and non-co-op sales. Carney does not cross-
classify type of brokerage by age of house, and it is clear from the
two-way tabulations that there are new homes in both categories in
most if not all of the samples, but none of the Žve samples with
fewer than 10 new homes have a signiŽcant dierence between
co-op and non-co-op commission rates. This pattern, combined with
the Goolsby and Childs result, suggests that the measured eect of
type of brokerage may in fact reect the dierence between new and
existing home sale arrangements.
   Commission rates are generally expected to vary inversely with
house price, on the basis that the eort needed to sell a home is not
proportional to the price of the house; although it is also argued that
the market for high-priced homes is smaller and therefore more ef-
fort may be needed to sell a home at or close to the seller’s reserva-
tion price.38 Carney Žnds lower rates on higher-priced homes in 14
of 19 samples, including six of the seven with statistically signiŽ-
cant dierences. (The exception is Jacksonville.) The other studies
all also Žnd negative relationships. The magnitude of the eect is
not clear, because the studies use dierent formulations, but it ap-
pears to be small. Goolsby and Childs Žnd that the commission rate
declines by about 0.06 to 0.11 percentage points for each $10,000
increase in home price, e.g. from 5.90 percent to 5.84 or 5.79
percent. Sirmans, Turnbull and Benjamin use the natural logarithm
of house price, and Žnd that the commission rate declines by about
0.04 percentage points per $10,000 increase in price around the
mean price of $92,000 in the sample, e.g. from 5.81 percent to 5.77
percent. Sirmans and Turnbull’s later study of Baton Rouge uses
both the price and the square of the price to test for non-linearities,
and Žnds that price declines at a decreasing rate. The net eect is
that commission rate declines by about 0.02 percentage points per

       Carney (1982) discusses these relationships.

132                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

$10,000, for sales prices around $100,000, e.g. from 6.00 percent to
5.98 percent. Carney simply reports correlation coecients between
home price and commission rate.
   The length of time that a home is on the market is important to
the seller as well as the sales price; indeed, ‘‘days on market’’ is one
of the research categories in the BJS bibliography. The paper by
Sirmans, Turnbull and Benjamin, which develops a three-equation
simultaneous model to explain house price, commission rate, and
days on market, Žnds a positive eect for days on market in the
commission rate equation; the longer it takes to sell the house, other
things being equal, the higher the commission rate. They also Žnd
that the higher the commission rate, other things being equal, the
longer it takes to sell the house. The papers by Goolsby and Childs,
and by Sirmans and Turnbull, both use single-equation models of
the commission rate, and Žnd no relationship with days on market.
Larsen and Park reverse the relationship, treating the commission
rate as an independent variable in an analysis of time on market
before sale. They Žnd a negative relationship between the size of the
commission and the probability that a property will sell at any given
time. Houses with lower commissions sell faster. This is the op-
posite result from Sirmans, Turnbull, and Benjamin.
   The two Baton Rouge studies relate the size of the brokerage Žrm
to the commission rate. The earlier paper Žnds no relationship, while
the later Žnds that large Žrms charge higher commission rates.
   The purpose of the Sirmans and Turnbull paper is to test a theo-
retical model relating commission rates to changes in the housing
market. They focus on the cyclical behavior of commission rates,
hypothesizing that rates rise during booms and fall during
recessions. Their results show that pattern: contract commission
rates rose during the early years, shortly after the collapse of oil
prices, and then began falling in the later 1980s. Also, contract com-
mission rates are negatively related to employment. The better the
economy, the lower the commission rate.39
   Thus to summarize the Žndings, commission rates appear to be
lower on higher-priced houses and lower on new homes. The evi-
dence is mixed on whether rates are lower when the listing and sell-

    In the model developed by Sirmans and Turnbull, this result is to be
expected because an increase in the demand for houses results in a greater
number of sales at higher prices and an increase in the number of brokers,
driving the commission rate lower. The possible osetting eect theoretically,
an increase in industry costs driving up costs for individual Žrms, is not strong
enough to generate the opposite empirical result.


ing broker are the same, and whether they are dierent for houses
that take longer to sell. The one study that looks at rates in a
macroeconomic context Žnds that they are higher in recessions and
lower in booms.

      Limitations of the Cross-Section Literature
   Unfortunately for the purposes of this paper, there are limitations
to most if not all of these studies which aect their value or
relevance for the purposes of the present paper.
   As noted, Larsen and Park do not try to explain the commission
rate; their interest is in explaining the number of days on the market.
While they report the full distribution of commission rates, they do
not attempt to explain why some rates are higher or lower than the 7
percent norm in Lincoln. Thus they have no Žndings to be included
in Table 7.
   The analysis by Goolsby and Childs suers from multicollinear-
ity problems. The selling broker’s share is the dependent variable
(termed R); one of the independent variables is the selling broker’s
commission multiplied by the list price (E}RL); and another is the
sales price (S). The variable E (commission x list price) is very
highly correlated with S (selling price); the simple correlation coef-
Žcient is y0.93. In the regression analysis, the following patterns
  (1) the coecients for E (itself a function of the dependent vari-
      able) have very high t-ratios, between 15 and 50, with posi-
      tive coecients;
  (2) the coecients for S, highly and positively correlated with
      E, also have very high t-ratios, between 15 and 40, but have
      the opposite signs from the coecients for E.
   Taken at face value, these results imply that the higher the sales
price, the lower the commission rate; but the higher the list price,
the higher the commission rate. This is possible, but unlikely. The
more logical explanation is that the high multicollinearity between
these variables has aected the results. A safer procedure would
have been to exclude the variable E.
   The two Baton Rouge studies are the most sophisticated. The
earlier paper by Sirmans, Turnbull, and Benjamin is the only one
with a simultaneous-equations model to explain the commission
rate; the later paper by Sirmans and Turnbull uses the most elabo-
rate econometric speciŽcations of any study, with several quadratic
terms to capture non-linearities, and with regressions employing

134                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

instrumental variables as well as ordinary least squares. It also has
much the largest data set, consisting of 15,608 home sales over the
eight years from 1985 through 1992. Unfortunately for my purposes,
the dependent variable in the later study is the contract commission
rate on the listing rather than the actual rate on the sale, and thus
‘‘does not reect any adjustments or changes that might be renegoti-
ated between the house seller and the agent at the time of sale’’.40
Given the FTC Žnding that some sellers enjoyed reductions in the
actual commission they paid, compared to the contract commission,
this casts some doubt on the validity of the Žndings. It is not clear if
the commission rate in the earlier study is also the contract
commission. If it is, the same problems arise.
   Carney uses the least sophisticated methodology. The simple
tabulations limit any inferences about the validity of the relation-
ships; as noted previously, it is possible that the higher commissions
on co-op sales reect dierences in the extent of co-op sales in the
markets for new and existing homes. Carney does perform multiple
regression analyses combining the data for all cities in a given year,
with results that match the simple correlations, but each of these
regressions includes cities with dierent modal commissions. In ef-
fect, he is trying to explain why commissions are typically 6 percent
in some markets and 7 percent in others, and at the same time why
some commissions are higher or lower than the mode in each
   A dierent approach to deŽning ‘‘commission rate’’ has been
taken in some research. Frew and Jud (1987) analyzed the relation-
ship between sales price and the use of the MLS through a broker,
using a sample of over 3,300 existing homes sold in Charlotte in
1977. They found that homes sold through the MLS had a higher
sales price by two to three percent. This result has been interpreted
to mean that the ‘‘net commission rate’’ is the relevant rate, and that
the net rate is subject to competition among brokers, even if the
gross commission rate is more or less Žxed.41 Other economists have
found the opposite result, however, and argued that the theoretical
relationship between MLS listing and price is indeterminate.42 What-
ever the empirical relationship, it seems clear that the ‘‘net commis-
sion rate’’ is not the relevant rate. The ability of brokers to provide
useful services to homebuyers and sellers is not at issue, and it is

       Sirmans and Turnbull (1991, p.111).
       Zumpano and Hooks (1988, p.13).
       Yavas and Colwell (1995).


reasonable to expect that both would pay for such services. That
does not imply that gross commission rates would be uniform and
rigid in a competitive market.

          Interpreting the Cross-Section Studies
   It is dicult to draw any very Žrm conclusions from such a small
body of empirical research. Some economists have suggested that
they may refute the common perception that commission rates are
rigid.43 That is a perfectly possible interpretation, and may be true.
But there are some reasons to think that such a conclusion is
premature, and some anomalies in the results that bear further
   (1) The popular view that all commissions are six percent (or
seven percent, or some other even number) is surely incorrect; there
is certainly variability, as shown by the mere existence of standard
deviations and regression results using the commission rate as the
dependent variable. But even when there is a high concentration of
rates at one number, the corresponding standard deviation can be
large enough to suggest more variability than the underlying data
actually show. In addition, of course, the standard deviation will be
aected by one or a few outliers in the distribution. The eect can
be illustrated by the FTC’s sample for Washington in 1979; one of
the 75 commission rates is below 5 percent, while all the others are
between 5 and 7 percent. When the one outlier is excluded, the stan-
dard deviation drops from 0.5 percent to 0.3 percent.44 This general
pattern can be interpreted either as an indication of a competitive
market, with a few outliers, or alternatively as giving a somewhat
misleading impression of the degree of dispersion and competition
that actually exists in the market.
   (2) What is surprising about these comparisons is that the earlier
data show the greater commission rate variability. The common
perception seems to be that commissions were more rigid in the
past. Multiple Listing Services often established commission rate
schedules that they recommended to their members, until the

    See for example Zumpano and Hooks (1988, p.13); Benjamin, Jud and
Sirmans (2000:2, p.18).
    The FTC does not report the precise rates for commissions other than the
even rates of 5, 6 and 7 percent; others are listed as ‘‘less than 5%,’’ ‘‘5%-
6%,’’ ‘‘6%-7%,’’ and ‘‘greater than 7%.’’ For this reason, I have used the
standard deviations reported by Carney, who had access to the data.

136                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

Department of Justice brought antitrust cases against the practice.45
NAR adopted a policy prohibiting members from establishing
recommended schedules in 1971, but it seems reasonable that habits
would change only gradually, and commission rates would show
more variability as time passed – more in the mid-1980s than the
later 1970s, for example.
   (3) Existing homes have generally carried higher commissions
than new ones, and this practice may still be common;46 thus a data
set containing both new and existing home transactions is likely to
indicate more dispersion than really occurs in the market. Similarly,
transactions involving both a listing and a selling broker may carry
larger commissions than transactions where one broker performs
both functions; the research results are mixed, but the dierence is
also supported by current anecdotes.47 A data set containing both
will therefore also show dispersion in commission rates. Further,
the data will show more dispersion, the more equal the split in the
data set between new and existing homes, or between co-op and
non-co-op transactions. Unfortunately, the published studies do not
provide enough information to evaluate this contention. Carney
reports standard deviations for the separate categories (existing
homes vs. new homes, co-op vs. non-co-op), but not for the full data
set, while Larsen and Park, Goolsby and Childs, and Sirmans, Turn-
bull and Benjamin, all report standard deviations for the full data set
but not the categories. One might expect smaller standard devia-
tions for Carney’s individual categories than for the full data sets in
the other two studies, but the opposite is generally the case.
   (4) Carney provides interesting information about California
commission rates. He reports that California historically (between
1935 and 1960) had a tapered rate structure, quoting a 1960 listing
by the San Francisco Board of six percent on the Žrst $25,000, Žve
percent on the next $75,000, and 2.5 percent above that. Carney fur-
ther states that such tapered structures are common ‘‘today’’ (pre-
sumably around 1980) in both the United States and England, with
citations to support his statement. If Carney is correct, the tapered
rate structure is a possible explanation for the size of the standard
deviations. In his data, the three Los Angeles city/year samples, and
the state of California, have fairly typical standard deviations: 1.0,
1.1. and 0.7 percent for Los Angeles, and 0.9 for California. A fairly

       GAO (2005, pp.12-13).
       Harney (2004), quoting anecdotal evidence.
       Harney (2004).


large standard deviation could arise even if all sales commissions
rigidly adhered to the same tapered commission rate schedule; it
would depend on the distribution of house prices. Moreover, if
tapered commission rate schedules were still widespread at the time
Carney wrote, but gradually became less common, then standard
deviations would tend to decline over time. This may help to explain
some of the dispersion in the literature.

Comparing the Time-Series and Cross-Section Data
   The cross-sectional literature suggests several possible explana-
tions for the apparent downward trend in the time-series data
produced by REAL Trends.
   The general cross-sectional Žnding that commission rates are
lower on higher-priced houses is certainly consistent with the
reported decline in rates. This hypothesis does not explain the rate
stability in very recent years, as house prices continued to rise, if
anything more sharply since 2001 than in the preceding decade.
   The Žnding that commission rates are lower on new homes has
less apparent relevance to the trend. New homes have constituted
about the same share of the market since 1995, ranging between
14.5 percent and 15.1 percent of the total. The new home share was
unusually low in 1991, at 13.6 percent, but this dierence is not
enough to aect the average commission rate. Moreover, it should
be remembered that the REAL Trends data come from the larger
brokers, and what is important is any change in the share of new
homes in their transactions, not in the overall market.
   The Sirmans and Turnbull Žnding that commissions were higher
for homes listed by large Žrms suggests that the REAL Trends data
may have an upward bias. But the commission rate in their study is
the contract rate, not the actual rate paid by the seller. Sellers
contracted to pay higher rates to larger Žrms, but may not have actu-
ally done so.
   Similarly, the Sirmans and Turnbull result that commission rates
were countercyclical in Baton Rouge from 1985 to 1992 is consis-
tent with the reported trend. The economy has been generally
expanding over the period covered by REAL Trends, so the national
commission rates reported by REAL Trends should have been
declining. Rates did not rise during the recent recession, however.
Also, the Sirmans and Turnbull model suggests that rates in 1991,
the earliest publicly available from REAL Trends, may have been
unusually high because of the recession in the early 1990s. The trend
itself may be spurious, as a secular phenomenon.

138                      REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

  But none of these possibilities should be pushed very far; they are
only speculations. It is important to emphasize that any serious at-
tempt to synthesize the cross-section and time-series evidence is
dubious if not heroic; all the cross-sectional data (except the last
year or two of Sirmans and Turnbull) is for years before REAL
Trends began collecting national data.

   This is perhaps a long paper to discuss a small body of data. By
limiting myself to systematic empirical studies, I have ignored a
much larger corpus of theoretical research on real estate brokerage.
Similarly, by trying to look systematically at commission rates over
time, I have limited myself to a single, proprietary, and private data
series, which is reported only sporadically in the press. There is not
much to talk about. This is largely because the data needed to talk
about it are very closely kept.
   There is perhaps one important inference that can be drawn from
the literature. The cross-sectional studies may provide a basis for
reconciling the common perception that commission rates are
administered and ‘‘sticky’’ with the research Žndings that show
variability and reasonable relationships between commission rates
and other phenomena. Commission rates for a given set of transac-
tions may cluster rather tightly around a norm, and still vary because
the norms are dierent for dierent kinds of transactions (e.g., exist-
ing vs. new homes), and because some brokers do lower some com-
mission rates to eectuate a sale. But it is not at all clear how much
the rate dispersion in these studies results from the variability in the
kinds of transactions and how much from actual variability in com-
mission rates.
   The most obvious conclusion is the academic economist’s stan-
dard complaint: we need further research. In this case, we Žrst need
the data, in order to do the research.
                                 Table 1
         Average Real Estate Commission from REAL Trends Survey

  Year       Average                          Source
 1988          —
 1989          —
 1990          —
 1991         6.1%                    Hahn, Litan and Gurman
 1992          —
 1993          —


   Year        Average                              Source
 1994             —
 1995           5.98%                              Lehmann
 1996             —
 1997             —
 1998            5.5%                U.S. Government Accountability Oce
 1999             —
 2000           5.48%                              Lehmann
 2001            5.4%                         Roberts and Mara
 2002            5.1%                         Roberts and Mara
 2003            5.1%                     Roberts and Mara; Lehmann
 2004            5.1%                      Hahn, Litan and Gurman;
                                     U.S. Government Accountability Oce
Robert W. Hahn, Robert E. Litan, and Jesse Gurman, ‘‘Paying Less for Real Estate
Brokerage: What Can Make It Happen?’’ Working Paper 05-11, August 2005, AEI-
Brookings Joint Center for Regulatory Studies
R.J. Lehmann, The 6% Solution, In Lehmann’s Terms, May 10, 2005.
Glenn Roberts, Jr. and Janis Mara, ‘‘Real Estate commission under pressure/Part 1:
ProŽts high despite commission erosion,’’ Inman News, April 18, 2005.
U.S. Government Accountability Oce, ‘‘REAL ESTATE BROKERAGE: Factors
That May Aect Price Competition,’’ GAO-05-947, August 2005.
                                Table 2
            Comparison of REAL Trends and Market Home Prices

 Year REAL Trends Existing Home      New Home      Existing    New      Weighted
      Average Home Average Price      Average      Homes     Homes      Average
          Price                         Price        Sold      Sold       Price
 2002   $229,530    $199,200          $228,700    5,631,000 973,000     $203,547
 2003   $247,721    $215,000          $246,300    6,183,000 1,086,000   $226,453

REAL Trends, Inc., ‘‘Analysis of the 2004 REAL Trends 500,’’ REAL Trends Monthly
Newsletter, April 2004.
U.S. Department of Housing and Urban Development, U.S. Housing Market Condi-
tions, 3rd Quarter 2005 (November 2005), Tables 6-9.
                                  Table 3
             Commission Rates by Firm Size and Region, 2002-2003

 Firm Category:             2002 Average               2003 Average
 Size of Firm               Commission                 Commission
 Top 50                     5.04%                      5.06%
 Top 100                    5.08%                      5.07%
 Top 101-500                5.26%                      5.24%
 Up-and-Comers              6.20%                      5.36%

 Firm Category: Region
 Northeast                  5.20%                      5.14%

140                        REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

 Firm Category:             2002 Average                    2003 Average
 Size of Firm               Commission                      Commission
 Middle Atlantic            4.78%                           4.78%
 Southeast                  5.50%                           5.38%
 Midwest                    5.62%                           5.38%
 Southwest/Mountain         5.16%                           5.26%
 Far West                   4.92%                           5.00%
REAL Trends, Inc., ‘‘Analysis of the 2004 REAL Trends 500,’’ REAL Trends Monthly
Newsletter, April 2004.
                                    Table 4
                          Average Commission per Sale

  Year    Average   Average Existing           Average                 Average
         Commission   Home Price              Commision              Commission
                                               Per Sale            in 2004 Dollars
 1988         —
 1989         —
 1990         —
 1991        6.1%          $128,400                $7,832             $10,862
 1992         —
 1993         —
 1994         —
 1995       5.98%          $139,100                $8,318             $10,310
 1996         —
 1997         —
 1998        5.5%          $159,100                $8,750             $10,140
 1999         —
 2000       5.48%          $176,200             $9,656                $10,592
 2001        5.4%          $185,300            $10,006                $10,672
 2002        5.1%          $199,200            $10,159                $10,667
 2003        5.1%          $215,000            $10,965                $11,257
 2004        5.1%          $236,600            $12,067                $12,067
Average Commission: Same as Table 1
Average Existing Home Price: National Association of Realtors
Consumer Price Index: Bureau of Labor Statistics.
                                    Table 5
                 Brokerage Commission Rates from FTC Study
                   (Calculated Rates on Existing Home Sales)

 Metropolitan Area       Mode         Percent At        Mean         Standard
 And Year                Rate         Mode              Rate         Deviation
 Bridgeport, 1975        6%           83%               5.7%         0.8%
 Washington, 1975        6            68                5.6          0.9
 Des Moines, 1975        7            80                6.9          0.4
 Denver, 1975            7            83                6.6          1.1
 Los Angeles, 1975       6            78                5.7          1.0


 Metropolitan Area        Mode           Percent At       Mean        Standard
 And Year                 Rate           Mode             Rate        Deviation
 Washington, 1978         6              72               6.0         0.9
 Houston, 1978            6              80               5.7         0.8
 Columbus, 1978           6              56               6.3         0.8
 Rochester, 1978          6              57               6.2         1.3
 Seattle, 1978            7              68               6.6         0.8
 Boston, 1979             6              53               6.2         1.1
 Denver, 1979             7              59               6.4         0.9
 Washington, 1979         6              84               5.9         0.5
 Jacksonville, 1979       7              30               6.6         1.2
 St. Louis, 1979          6              85               5.8         0.8
 San Antonio, 1979        6              82               5.6         1.0
 Seattle, 1979            7              57               6.4         0.9
 Boston, 1975*            6              56               5.6         1.2
 Orlando, 1975*           7              51               6.5         1.1
 Los Angeles, 1978*       6              78               5.7         1.1
 Atlanta, 1978*           7              69               6.6         1.2
 Portland, 1978*          7              56               6.3         1.5
 Los Angeles, 1979*       6              80               5.9         0.7

  * Sample size less than 50
Mode and percent at mode: Federal Trade Commission, ‘‘The Residential Real Estate
Brokerage Industry: Volume 1,’’ p. 51, Table III-6
Mean and standard deviation: Carney, ‘‘Costs and Pricing of Home Brokerage Ser-
vices,’’ p. 340, Exhibit 4
                                  Table 6
              Brokerage Commission Rates from Academic Studies
                           of Individual Markets

 Authors            Metropolitan         Mode Rate Percent At    Mean     Standard
                    Area and Year                  Mode          Rate     Deviation
 Larsen and Park*   Lincoln, 1986        7%        84%           6.8%     0.5%
 Goolsby and Childs Knoxville,           Not       Not           5.9%     0.8%
                    1983 & 1987          Reported Reported
 Sirmans, Turnbull Baton Rouge,          Not       Not           5.8%     0.6%
 And Benjamin       1985-1987            Reported Reported
  * Data limited to home actually sold
                                     Table 7
                       Factors Aecting Commission Rates

 Authors         Higher        New             Co-op       Longer        Large Žrm?
                 Price For     Home?           Sale?       Time On
                 Home?                                     Market?
 Goolsby &       Lower rate    Lower rate      No eect    No eect      Not studied
 Sirmans,        Lower rate    Not studied Not studied Higher rate No eect
 Turnbull, and

142                         REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

Authors         Higher        New          Co-op       Longer         Large Žrm?
                Price For     Home?        Sale?       Time On
                Home?                                  Market?
Sirmans &       Lower rate    Lower rate   Not studied No eect       Higher rate
Carney (FTC)    Lower         Lower        Higher       Not studied Not studied
                rate* (6      rate** (10   rate** (8 of
                lower, 1      of 10)       16)
                higher of
 * Limited to areas with sample size of at least 50 homes
 ** Limited to areas with sample size of at least 20 homes in each category

Paul Anglin and Richard Arnott, ‘‘Are Brokers’ Commission Rates
on Home Sales Too High? A Conceptual Analysis,’’ Real Estate
Economics, Volume 27, No. 4 (1999), pp. 719-749.
Anna Bahney, ‘‘Are Commissions Really Negotiable? Sort Of,’’
New York Times, May 15, 2005.
John D. Benjamin, G. Donald Jud, and G. Stacy Sirmans, ‘‘Real
Estate Brokerage and the Housing Market: An Annotated Bibliogra-
phy,’’ Journal of Real Estate Research, Volume 20, No. 1-2 (2000),
pp. 217-278.
John D. Benjamin, G. Donald Jud, and G. Stacy Sirmans, ‘‘What
Do We Know About Real Estate Brokerage?’’ Journal of Real
Estate Research, Volume 20, No. 1-2 (2000), pp. 5-30.
Donald Bruce and Rudy Santore, ‘‘A Note on Optimal Real Estate
Commissions,’’ unpublished paper, September 2004.
Michael Carney, ‘‘Costs and Pricing of Home Brokerage Services,’’
AREUEA Journal, Volume 10, No. 3 (1982), pp. 331-354.
Natalya Delcoure & Norm G. Miller, International Residential Real
Estate Brokerage Fees and Implications for the US Brokerage
Industry, 5 Int’l Real Estate Rev. 12 (2002).
Blanche Evans, ‘‘NAR Membership Grows: Are New Members
Getting Younger?’’ Realty Times, February 11, 2003.
Blanche Evans, ‘‘Are The Media Attacks on Realtors Eective?’’
Realty Times, August 25, 2005.
James R. Frew and G. Donald Jud, ‘‘Who Pays the Real Estate
Broker’s Commission?’’ Research in Law and Economics, Volume
10, pp. 177-187.
William C. Goolsby and Barbara J. Childs, ‘‘Brokerage Firm Com-
petition in Real Estate Commission Rates,’’ Journal of Real Estate
Research, Volume 3, No. 2 (1988), pp. 79-85.

REFERENCES                                                    143

James R. Hagerty, ‘‘Crowded House: Discount Real-Estate Brokers
Spark a War Over Commissions,’’ Wall Street Journal, October 12,
2005, p. A1.
Robert W. Hahn, Robert E. Litan, and Jesse Gurman, ‘‘Paying Less
for Real Estate Brokerage: What Can Make It Happen?’’ Working
Paper 05-11, August 2005, AEI-Brookings Joint Center for Regula-
tory Studies.
Ken Harney, ‘‘What Sellers Really Pay in Commission,’’ TheRe-
alDeal, August 2004
Kannon Consulting, ‘‘2004 Horizon Watching Initiative, Board
Committee on Industry Development, Real Estate Baseline, July 23,
2004,’’ presentation to the Newspaper Association of America.
Robert H. Lande & Howard P. Marvel, The Three Types of Collu-
sion: Fixing Prices, Rivals, and Rules, 2000 Wisconsin Law Review
James E. Larsen and Won J. Park, ‘‘Non-Uniform Percentage
Brokerage Commissions and Real Estate Market Performance,’’
AREUEA Journal, Volume 17, No. 4 (1989), pp. 422-438.
David Lereah, ‘‘As competitive as it gets,’’ Realtor Magazine On-
line, September 1, 2005.
R.J. Lehmann, ‘‘The 6% Solution,’’ In Lehmann’s Terms, May 10,
REAL Trends, Inc., ‘‘Analysis of the 2004 REAL Trends 500,’’
REAL Trends Monthly Newsletter, April 2004.
( –
Alan Murray, ‘‘Brokers Fiddle as Real Estate Burns,’’ Wall Street
Journal, September 28, 2005, p. A2.
Clay Risen, ‘‘Home Page,’’ New Republic, May 2 & May 9, 2005.
Glenn Roberts, Jr. and Janis Mara, ‘‘Real Estate commission under
pressure/Part 1: ProŽts high despite commission erosion,’’ Inman
News, April 18, 2005.

144                    REAL ESTATE LAW JOURNAL [VOL. 35: 119 2006]

C.F. Sirmans and Georey K. Turnbull, ‘‘Brokerage Pricing Under
Competition,’’ Journal of Urban Economics, Vol. 41 (1997), pp.
C.F. Sirmans, Georey K. Turnbull, and John D. Benjamin, ‘‘The
markets for housing and real estate broker services,’’ Journal of
Housing Economics, Volume 1, No. 3 (1991), pp. 207-217.
U.S. Bureau of Labor Statistics, Current Employment Statistics
survey, Series CEU5553120001 and CEU5553120003. (http://
U.S. Bureau of Labor Statistics, Current Population Statistics
survey, (, Table 11 for current
year; ( for
years since 2000;
aat11.txt for earlier years back to 1995.
U.S. Department of Housing and Urban Development, U.S. Hous-
ing Market Conditions, 3rd Quarter 2005 (November 2005).
U.S. Department of Justice, ‘‘Justice Department Sues National As-
sociation of Realtors for Limiting Competition Among Real Estate
Brokers,’’ press release, September 8, 2005.
U.S. Federal Trade Commission, The Residential Real Estate
Brokerage Industry: Volume 1, Washington: U.S. Government
Printing Oce, December 1983.
U.S. Government Accountability Oce, ‘‘REAL ESTATE BRO-
KERAGE: Factors That May Aect Price Competition,’’ GAO-05-
947, August 2005.
Wall Street Journal, ‘‘The Realtor Racket,’’ editorial, August 12,
2005, p. A8.
Abdullah Yavas and Peter F. Colwell, ‘‘A Comparison of Real
Estate Marketing Systems: Theory and Evidence,’’ Journal of Real
Estate Research, Volume 10, No. 5 (1995), pp. 583-599.
Leonard V. Zumpano and Donald L. Hooks, ‘‘The Real Estate
Brokerage Market: A Critical Reevaluation,’’ AREUEA Journal,
Volume 16, No. 1 (1988), pp. 1-13.
Leonard V. Zumpano, Ken H. Johnson, and Randy J. Anderson,
‘‘Internet use and real estate brokerage market intermediation,’’
Journal of Housing Economics, Volume 12 (2003), pp. 134-150.


Shared By:
Description: Real Estate Brokerage Financial Reports document sample