MLS versus - Society for Economic Dynamics

MLS versus - Society for Economic Dynamics

The Relative Performance of Real Estate Marketing Platforms: MLS versus Igal Hendel Aviv Nevo François Ortalo-Magné June 7, 2007 A...

143KB Sizes 0 Downloads 0 Views

Recommend Documents

edinburgh - Society for Economic Dynamics
University of Edinburgh, central to this ... Photography: Neale Smith (The University of Edinburgh Old College Quad), Pa

Schumpeterian Entrepreneurship - Society for Economic Dynamics
Feb 14, 2009 - Schumpeterian entrepreneurship in which founders exploit ideas they learned ... Perhaps the best known is

Korean Housing Repo Contract - Society for Economic Dynamics
Feb 11, 2013 - development. Koreaes unique Jeonse or housing repo contract chan% nels funds directly from tenant/lenders

New Terrain in International Trade - Society for Economic Dynamics
Jun 27, 2015 - Bridging models: Arkolakis, Costinot and Rodríguez-Clare (ACR 2012). • Estimating the trade elasticity

Monetary Policy and Durable Goods - Society for Economic Dynamics
Mar 26, 2015 - We analyze monetary policy in a New Keynesian model with durable and non-durable goods, each with a separ

Inequality and Aggregate Demand - Society for Economic Dynamics
Inequality and Aggregate Demand. Adrien Auclert. Matthew Rognlie*. February 2016. Extremely preliminary. Abstract. We ex

Financial Risk and Unemployment - Society for Economic Dynamics
Financial Risk and Unemployment. Zvi Eckstein. Tel Aviv University and The Interdisciplinary Center Herzliya. Ofer Setty

Land Reform and Productivity - Society for Economic Dynamics
farmer over time and therefore to observe the source of the within-farm effect following the reform, such as a change in

Complexity in Structured Finance - Society for Economic Dynamics
Sep 13, 2013 - Financial Wizardry or Smoke and Mirrors?˚ ... ask what role, if any, complexity plays in structured fina

Not-Quite-Great Depressions of Turkey - Society for Economic Dynamics
Feb 10, 2009 - Turkish economy. Keywords: growth accounting, total factor productivity, great depressions, Turkey, dy- n

The Relative Performance of Real Estate Marketing Platforms: MLS versus Igal Hendel

Aviv Nevo

François Ortalo-Magné

June 7, 2007

Abstract A good real estate agent might make up some of the commision he or she is paid by helping the seller get a more favorable outcome. We match several data sets to compare the outcomes obtained by sellers who listed their home on a For-Sale-By-Owner (FSBO) web site versus those who used an agent and the Multiple Listing Service (MLS). We do not …nd that listing on the MLS helps sellers obtain a signi…cantly higher sale price. Listing on the MLS does shorten the time it takes to sell a house.

We are grateful to the owners of and the South-Central Wisconsin Realtors Association for providing us with their listing data. Geo¤ Ihle and James Robert provided valuable research assistance. Igal Hendel and Aviv Nevo thank the Center for the Study of Industrial Organization and the Guthrie Center for Real Estate Research at Northwestern University. François Ortalo-Magné acknowledges …nancial support from the James A. Graaskamp Center for Real Estate and the Graduate School at the University of Wisconsin–Madison. We bene…ted from the comments of Morris Davis and seminar participants at the University of Chicago, Duke University, Harvard University, MIT, Northwestern University, Stanford University, the University of Toronto, the University of Wisconsin-Madison, Yale University. Igal Hendel and Aviv Nevo are in the department of Economics at Northwestern University and NBER. François OrtaloMagné is in the department of Economics and the department of Real Estate and Urban Land Economics at the University of Wisconsin-Madison. Contact information: [email protected], [email protected], and [email protected]




A large number of housing transactions are carried out with the help of realtors.1 Realtors provide expertise (on pricing, conditioning the house for sale and bargaining) and convenience (by showing the house, advertising and holding open houses and helping with the paperwork). Another advantage of working with a realtor is access to the Multiple Listing Service (MLS), a database that compiles information on all the properties listed by local realtors. For these services realtors charge a commission at, or around, 6%. The commission rate has been stable over time and across regions and has been the subject of scrutiny antitrust authorities. The advent of the internet has a¤ected many markets. The real estate market is one of them. Internet sites facilitate direct (by owner) marketing. Direct marketing was always possible using newspapers, ‡yers and other forms of advertising. However, the internet is a cheaper and potentially more e¤ective form of direct marketing. Sellers can post rich information, photos as well as virtual tours. For-Sale-By-Owner (FSBO) web sites provide an alternative platform, or two-sided network, that competes directly with the MLS network. In this paper we study the functioning of intermediation, or brokerage, in real estate markets. Brokerage in real estate markets is important for several reasons. According to the Bureau of Economic Analysis the size of the real estate brokerage market as a share of GDP has been increasing over time, and in 2005 accounted for 0.9% of GDP. In addition to being a large industry, e¢ ciency in the brokerage market impacts the functioning of real estate markets. Finally, from a public policy point of view, foreclosure of competition from the MLS has recently been the focus of several antitrust cases.2 We assess the role and performance of realtors relative to sales by owner. In particular, we quantify the cost of hiring a realtor. The actual cost of hiring a realtor is the commission minus the price premium that the agent might generate and the …nancial savings from a faster sale. The price premium may largely o¤set the commission, or even more than make up the commission.3 We quantify the actual monetary cost of using an agent by comparing the 1

Real estate agents are licensed by the state. A realtor is a real estate agent who is a member of the Realtor Association. 2 See for example or 2006/10/realestatesweep.htm. 3 The National Association of Realtors website claims, based on the 2005 Home Buyer & Seller Survey


performance of listing by owner to transactions with realtors. We also assess other outcome measures like time on the market and the overall probability of sale for these alternative channels. We focus on the city of Madison, Wisconsin, where a single web site ( has become the dominant for-sale-by-owner platform. With the cooperation of we gained access to all FSBO listings since the start of the platform. We combined the FSBO data with data from two other sources. First, from the South-Central Wisconsin Realtors Association we got access to all MLS listings in the city. Second, we matched every listing with data from the city of Madison. The city of Madison assessor o¢ ce maintains a database with the full history of transactions on every property together with an exhaustive set of property characteristics. By merging these data sets we get a complete history of events that occurred for virtually every single family home listed in the city between January 1998 and December 2004. A history of a listing includes: date and platform of initial listing, moves across platforms, and outcome (eventual sale date, and price if sold). These data sets allow us to compare the performance of the listing platforms in terms of marketing outcomes. We …nd that the average sale price of homes that sell on FSBO is higher than the average price of homes that sell with a realtor. The characteristics, reported in the city assessor’s database, of houses sold on the di¤erent platforms are somewhat di¤erent. However, after controlling for the these observed characteristics the FSBO premium persists. Even with the very rich set of controls at our disposal two concerns remain. First, there might be unobserved house characteristics that a¤ect both the decision to sell on FSBO and prices. For example, homes that are easier to sell (i.e., conform better to the taste of the population) may be more likely to be listed and sold through FSBO. At the same time these popular homes may confer a price premium. To deal with unobserved house heterogeneity we examine properties that sold multiple times. Estimates are essentially identical to those computed using just a single sale and a rich set of controls. We therefore conclude that unobserved house heterogeneity, that is …xed over time, does not seem to be a problem. that "the median home price for sellers who use an agent is 16.0 percent higher than a home sold directly by an owner; $230,000 vs. $198,200; there were no signi…cant di¤erences between the types of homes sold."


The second concern is the selection of sellers into FSBO. Sellers may di¤er, for example, in their patience or bargaining ability.4 More patient sellers are likely to get a better price, regardless of the platform they choose. At the same time they may be more prone to list on FSBO. In that case we will get a positive correlation between FSBO and sale price. We deal with the potential seller selection issue in several ways. First, we compare the houses that initially listed on FSBO, did not sell, but instead were eventually sold through MLS, to those that listed and sold on FSBO. These two groups of houses sell on di¤erent platforms but belong to the initial population that selected FSBO. If we think that the owners of these houses are similar, and that the reason some sold while others did not is luck of the draw, then the di¤erence in price will give us the causal e¤ect of FSBO. We …nd that houses that listed and sold on FSBO sell for a small, and not statistically signi…cant, premium compared to houses that listed on FSBO initially but that were eventually sold on MLS. Even if moving from FSBO to MLS depends on seller type the selection bias should be reduced, as the group of FSBO listers is more homogenous than the population as a whole. This comparison should at least provide a cleaner, perhaps not completely clean, platform comparison. Our second approach to deal with seller heterogeneity is to compare FSBO sales to realtors’ own properties sold on the MLS. Levitt and Syverson (2006) …nd a premium for realtors’ own properties sold on the MLS. They attribute this price gap to an incentive problem. Repeating the analysis in our data we get a premium almost identical to Levitt and Syverson. We compare this to the premium sellers get on FSBO. Both are by owner transactions, thus, do not su¤er from the agency problem identi…ed by Levitt and Syverson. Since realtors are professional this comparison should bound the impact of selection. Even if the homeowners who use FSBO are better bargainers than the typical homeowner, it is reasonable to assume they are no better at bargaining than professional realtors. We …nd that the FSBO premium is similar to the premium realtors obtain when selling their own homes. In line with the previous …ndings, this suggests no price di¤erences across platforms. The third approach we take to deal with seller heterogeneity is to compare transactions of 4

For a descriptive study of bargaining patters using English data see Merlo and Ortalo-Magné (2004), and Merlo, Ortalo-Magné and Rust (2006) for a structural model of bargaining using the same data.


the same seller using di¤erent platforms. We matched seller names across transactions and compare their performance across platforms. We …nd no price premium across platforms. Namely, the initial FSBO premium vanishes once we add a seller …xed e¤ect. To con…rm that the FSBO premium is explained by seller selection, we estimate the price premium of FSBO sellers while selling on the MLS. We de…ne as a FSBO seller those sellers that sell on FSBO sometime during the sample. Then we estimate the hedonic price regression for MLS transactions only. The FSBO seller dummy carries a premium similar to the FSBO premium. The estimate suggest the latter was driven by seller e¤ects rather than platform e¤ects. Finally, we examine various factors that impact the sellers decision to sell on FSBO as instrumental variables. For example, we used the fraction of previous sales on FSBO in the seller’s neighborhood. The point estimates we …nd are consistent with a FSBO premium. However, the instruments are very weak and the standard errors are very large. All the approaches used to deal with selection lead us to the same conclusion: the two platforms deliver the same prices. There is no support in our data to the claim that the MLS delivers a higher price. This is not to say that realtors do not provide value to the seller. The cost of such convenience provided by realtors seems to be the full commission (or half the commission since part of the FSBO transactions involve a realtor). Comparing other outcomes, we …nd that houses initially listed on FSBO tend to take slightly longer to sell but have a higher probability of eventually selling. The longer time to sell is driven by two factors. First, a proportion (about 20%) of FSBO listings that move to the MLS after initial failure. The shift from FSBO to MLS entails the risk of staying 68 more days on the market. Second, the probability of a quick sell is larger for houses initially listed on the MLS.


Realtors and

Historically, most real estate transactions have been performed using real estate agents. A homeowner wishing to sell their home will contract with a real estate agent o¤ering them exclusivity for a limited period, usually 6 months, and agreeing to pay a commission, of 5

usually 6% of the sale price, if the house is sold during that period.5 The commission is typically split between the listing agent, who is the agent that contracted with the seller, and the selling agent, who is the agent that brings the buyer. Some states, for example, Wisconsin, where our data come from, also recognize the status of buyer agency.6 If a buyer agent is involved in the transaction, s/he deals with the listing agent to settle the terms of the transactions, and gets the selling agent commission. When the same agent lists and sells the property, this agent gets the whole commission. In order to become a real estate agent one has to be licensed by the state. In most states this requires a short course and to pass a licensing exam. A real estate agent becomes a realtor when s/he joins the realtor association and subscribes to its code of ethics. Joining the association provides the agent with several advantages, one of them is access to the MLS. Working with an agent, and agreeing to pay the commission, gives the homeowner access to a number of services. The National Association of Realtors (NAR) argues that Realtors provide valuable help with setting the listing price, preparing the house, checking potential buyers’ quali…cations, showing the house, bargaining the terms of the deal, and handling the paperwork. Another advantage of working with a realtor is access to the MLS. In the market we examine this involves the ability to list on the South Central Wisconsin MLS, which costs a minimal fee, $10 as of 2007, but requires membership in the organization. In 1998 an alternative to the MLS was launched in Madison, Wisconsin: the website Christie Miller and Mary Clare Murphy recruited 9 listings from forsale ads in the local newspaper, added Mrs. Murphy’s house and launched their website with 10 listings. From the get-go, the strategy of was to provide a cheap no-frills service. In exchange for a fee of $75 initially, $150 for most of the period of our sample, homeowners can post their listing on the website (property characteristics, contact details and a few pictures). FSBO provides sellers with a yard sign similar to those provided by realtors but with the distinctive logo and color of Listings are kept active for 6 months, more if the fee is paid again. has been successful, 5

For a discussion of the commissions charged by agents see Hsieh and Moretti (2003) and references therein. 6 The di¤erence between a buyer agent and a selling agent is mostly a legal one having to do with the contractual agreement, or lack of it, between buyer and agent.


establishing itself as basically the only website for for-sale-by-owner properties in the city. Properties are removed from the site upon instruction of the homeowners. Typical events that trigger removal include sale of the property, withdrawal of the property from the market, or transfer of the property to the MLS platform. The sta¤ of monitors listings on the MLS and extinguishes any listing from their website that ends up on the MLS. This is done primarily to avoid disputes with the MLS. Real estate agents will occasionally be involved in FSBO sales when they represent the buyer and one of the parties to the transaction accepts to pay a buying agent commission, typically 3%. When such sales occur, the real estate agent may create a listing on the MLS and declare it as sold right away. In Madison, all such listings get a speci…c code that identi…es them as FSBO listings. This enables us to identify some of the FSBO sales that are executed with the help of a realtor without being listed by a realtor. Note that the typical buyer agency agreement does not allow the household to buy a FSBO home without payment of a commission to the realtor. Recently, a number of limited-service brokers have emerged. In Madison, the dominant …rm appears to be Madcity Homes ( Madcity Homes charges $399 to list a house on the MLS for 6 months and also provides the seller with a yard sign. The homeowner gets no other service. Additional services are available for an extra fee upon request. The homeowner is responsible for paying the 3% commission to any realtor that sells the house, whether the realtor is under buyer agency agreement or not. No commission must be paid if the sale does not involve a realtor. By the end of 2004, when our sample ends, this …rm had too few listings for us to analyze the extent to which limited-service brokerage yields di¤erent outcomes than full-service MLS listings or listings.


Theoretical Framework

In this section we brie‡y present a theoretical framework to think about the matching of buyers and sellers in the real estate market. Coles and Muthoo (1998) present a stock-and-


‡ow model of matching between unemployed workers and vacancies.7 Their stock-and-‡ow model, mildly adapted, will be useful to think about platform choice and selection issues. There are many issues like incomplete information, learning about market conditions or own property, that a¤ect decisions but we will not consider. The basic idea of their model is as follows. There is a ‡ow of new buyers (sellers) into the market in every period. Entrants are immediately put in contact with the stock of agents on the other side of the market. There is a probability

that a house …ts the needs of the buyer.

Buyers costlessly observe whether they have gains from trade with each house currently on sale. Namely, they …nd out which of the houses currently in the stock of houses for sale meet their needs. If they …nd a single agent to trade with, they split the gains from trade. If instead a newcomer meets multiple counterparts, she receives simultaneous o¤ers generating a Bertrand-type game. Agents that trade leave the market. Incoming buyers (sellers) that do not …nd a match, or fail to trade, join the stock of buyers (sellers). Coles and Muthoo show that in equilibrium matched players always trade (due to complete information). In equilibrium there is no trade among the stocks, if there were gains from trade they would have traded already. Thus, in equilibrium newcomers trade with the stock. The stock of buyers (sellers) only …nds gains from trade –match– with the ‡ow of sellers (buyers). We explore two variations: (i) we consider the coexistence of two competing platforms, F and M , where agents can participate and (ii) house and seller heterogeneity. The later will help us think about unobserved heterogeneity and potential biases once we get to the data. Platform Choice We make the following assumptions in order to capture the main practical di¤erences across platforms. First, we assume that the existence of the platform F is known to only a proportion of agents.8 Only informed agents have a choice, uninformed 7

See also Coles and Smith (1998), and Taylor (1995), and for a discussion of brokerage choice Salant (1991), Yavas and Colwell (1999) and Munneke and Yavas (2001). 8 Heterogeneity in the disutility of trading without a realtor can also drive platform choice. Some sellers are aware of the option of sale by owner but may …nd it too costly to show the house and bargain.


ones trade in M:9 Second, we assume there is an asymmetry between buyers and sellers. While informed buyers can shop on both platforms, sellers choose a single platform. This exclusivity is required by the MLS. Third, listing in M , in addition to the exclusivity, involves a commitment to pay a transaction cost (or commission) C should the house sell within periods of listing. These assumptions make F a cheaper alternative, involves no fees. At the same time F involves less exposure, thus a lower matching rate. Heterogeneity We think of houses di¤ering in their degree of liquidity, . Owners of more liquid houses, which get more matches, may systematically opt for one of the platforms, and at the same time sell at a premium (as they generate more o¤ers). Sellers may also be heterogeneous, for example, in their patience or bargaining ability. Patience in this model will a¤ect both platform choice as well as transaction price given a platform. Implications Within this framework, informed buyers shop, and match, on both platforms. The probability of matching in either platform depends on sellers behavior, namely, on what proportion of the properties lists on each platform. Uninformed buyers and sellers face no choice, they shop exclusively on M: Informed sellers have to chose an exclusive platform. The trade o¤ is between an expensive and more e¤ective platform, M; and the non-fee F platform that o¤ers exposure to fewer buyers. For any speci…c property, the extra exposure leads to higher success rate. Claim 1 For given seller and house characteristics, on M we should observe shorter time to sell and higher success rate, holding time on the market …xed. The bene…t of listing in F is common to all sellers, however, the more patient the seller or liquid the property the less costly it is to use F: Thus, the appeal of F depends on seller patience and liquidity of the property, : Two implications are immediate. First, impatient sellers and non-liquid properties list in M . Moreover, they have no incentive to ever move to F should they fail to match in M . The reason is that buyers in F also shop in M; failure to match in M means that no matches 9

Although not necessary, it is reasonable to assume that the set of buyers aware of F is a subset of those aware of M. For example, out of town buyers are less likely to be familiar with fsbomadison.


will be found in F either. Having explored all the stock of buyers, the seller can only wait for the ‡ow of incoming buyers. Since the ‡ow is larger in M; impatient sellers stay there. In contrast, patient sellers and owners of liquid properties prefer to list in F: If they fail to match in F; they move to M to try to match with the rest of the stock of buyers (those that shop only on M ): Once they explored M , all stock has been exhausted, thus, they have no incentive to move back to F: The incentives just described can be summarized in the following claims. Claim 2 A proportion of sellers try F …rst, if they fail to match they move to M and stay (matching the ‡ow in M). There are no moves from M to F

Claim 3 More patient sellers and sellers with easier to sell houses list on F …rst. F provides a cheaper way to explore a subset of the stock of buyers. The attraction of this option increases with the proportion of informed buyers, and declines with the number of sellers that list in F (sellers compete for the stock of buyers): As the number of informed buyers increases the success rate (probability a seller …nds a match) increases. However, the extra success draws more listings. As more informed buyers shop in F more sellers list, equilibrating the success rate. Claim 4 As the proportion of informed buyers increases the success rate at F is stable Since, given similar terms, buyers are indi¤erent between the platforms, as frictions disappear they would not pay any of the premium. Claim 5 As frictions vanish (i.e., more buyers become patient and informed about F) prices across platforms tend to coincide In sum, the model suggests that sellers are likely to list using FSBO to expose their property to a subset of the stock of buyers, if they fail to match, they move on the MLS for exposure to the rest of the stock, and subsequent ‡ow of buyers.




We obtained data from, the South-Central Wisconsin Realtors Association, the City of Madison and Dane County. We merged the date into a single database, organized by parcel numbers as designated by the City. MLS data The South-Central Wisconsin Realtors Association provided us with all listing activity on their Multiple Listing Service between 1/1/1998 and 5/23/2005. For each listing, we know the address of the property, its parcel number, whether the property is a condo or not, the listing date, and the status of the listing. In addition, whenever relevant, each record contains the expiration date of the listing, the accepted o¤er date, the closing date and the sale price as recorder by realtors. FSBO data The owners of the website provided us with information on all the listings with their service since it started in 1998. For each listing, we know the address of the property, the last name of the seller, the date the property is put on the web and sometimes information about the outcome of the listing. At this point, we use data for the years 1998-2004, with an address in Madison. City Data The city of Madison is located within Dane County. The city database provides information on sale prices and large set of property characteristics, about both the parcel and the buildings. In addition, the county maintains a county-wide database with location information for each parcel. We use this database to obtain spatial coordinates for each property. The county and the city do not use the same parcel numbers for condominium. Whenever there are such incompatibilities, we use Streetmap to locate the properties.


Descriptive Statistics

Merging the above data sets and excluding listings as de…ned above we get 15,616 listings, which represent 12,384 unique properties, in 1998 to 2004. In Table 1 we describe these listings. A row represents where the property was initially listed. The columns represent the eventual outcome of the listing, namely, whether it sold and how. Actual histories can be 11

more complicated, like listing with several agents or leave a platform to then return, but we mostly abstract from these complications. The market share of FSBO in listings during the entire sample period is roughly 20%. We de…ne a non-sale as any listing that showed up in either MLS or FSBO but was not recorded later in the city data with a sales price. Approximately 87% of the properties eventually sell. Out of the properties that sell, 95% sell through the initial listing platform. The remaining 5% are almost completely switches from FSBO to MLS. Switches from MLS to FSBO are almost nonexistent, accounting for just 0.3% of the MLS listings. This is consistent with the predictions of the model (i.e., Claim 1) by which some sellers may try the cheaper platform …rst but they have no incentive to return. Moreover, should they prefer not to try the stock in F they would not come back for its ‡ow. The market share of FSBO in properties sold is roughly 14%, slightly below its listing share. Since FSBO was only introduced in 1998, these numbers somewhat underestimate the current FSBO market share. Therefore, in the rest of Table 1 we present the breakdown for every other year of the sample. FSBO’s share in listing and in outcome increases over time. By 2004, the last year of the sample, FSBO share in listing is over 27%, and the share in sales is almost 20%. To judge the success of each platform we look at the proportion of properties that sell through their …rst listing. Of the 3,140 initial FSBO listings 2,153 or 68.6% sell on FSBO while 84.9% of initial MLS listings (10,718 out of 12,476) sell on MLS. While there is a clear trend in FSBO listing, increasing from 6% in 1998 to 27% in 2004, the trend in success rate is less clear. The success rate in 2004, 71.2%, is higher than the rate in 1998, 63.1%. However, there is no clear trend in the intermediate years. This is line with Claim 4. Just as the penetration of FSBO increases over time it also di¤ers across neighborhoods. In Table 2 we present the FSBO penetration rate across di¤erent assessment areas. These areas are de…ned by the City of Madison for assessment purposes. We get similar variation if we look at elementary schools areas. The FSBO listing share vary between 7.9% and 43.6.% The top FSBO share neighborhoods tend to be close to campus. Similar variation is present also in the FSBO share of sales. The success rate of FSBO listings varies by neighborhood. For neighborhood with at 12

least ten FSBO listings the success rate ranges from 31% to 100% (with one outlier at 9%). The mean success rate is 66% and the standard deviation is 13.2%. There is a positive relation between the propensity to list using FSBO and the success rate, which can be seen through a linear regression. Using the estimated slope, one standard deviation increase in the success rate translates into 2 percentage points increase in the propensity to list FSBO. In the analysis below we compare the performance of properties sold through FSBO and through MLS. A key question is whether these properties are comparable. In Table 3 we explore this issue. It compares several of the house characteristics in the data. The columns present the mean and standard deviation for properties listed initially through FSBO and MLS. The last two columns present the di¤erence between these means and the t-statistic of the di¤erence. The di¤erences in the means for most characteristics are small. However, because of the reasonably large sample sizes the di¤erences are signi…cant in some cases. For example, FSBO properties are somewhat older, tend to be on smaller lots and have smaller basements, but have somewhat newer roofs and furnaces.




Outcomes by FSBO and MLS channels

We now explore the di¤erences in outcome for properties sold through FSBO and MLS. In Tables 4-6 we present the results from regressing sale price, time on the market and the probability of a sale, on a FSBO dummy variable and various controls. In Table 4 we display the e¤ect of channel on price. In the top panel of the table the dependent variable is the logarithm of price, while in the bottom panel we regress the price level on various controls. The sample in columns (i) through (iv) includes only properties that sold on the channel they were originally listed. In the …rst column we regress price on a dummy variable that equals one if the house was sold on FSBO (divided by 100). If listing channel is determined at random, and the seller cannot switch from the channel they were assigned then this regression measures the causal e¤ect of selling on FSBO. In the spirit of this ideal situation the sample includes only houses that sold on the channel


they originally list. The results suggest that on average there is a large positive premium for selling on FSBO, roughly a 11 percent premium or 14,800 dollars. Since the dependent variable is the sale price, and not the sale price net of commission, this premium is on top of the saved commission. The magnitude of the premium is driven by the time trends in the data that we saw in Table 1. Over time prices have gone up and so has the share of FSBO sales. Indeed, once we control for year and month time dummy variables and a linear time trend, in column (ii), the e¤ect goes down to roughly 4 percent, or 3,000 dollars, but is still statistically signi…cant. The numbers in Table 3 suggest that there is some di¤erence in the observed characteristics of houses sold through FSBO and MLS. If the houses sold on FSBO have more attractive characteristics, then the FSBO dummy variable will also capture the impact of these features, rather than the e¤ect of selling through FSBO. Furthermore, Table 2 suggests that FSBO has a higher share in some areas. If these areas are more attractive this will bias our estimates. In order to control for the di¤erences in houses we construct an hedonic model of prices. Column (iii) reports the results from this model. In the controls we include the characteristics of the house, displayed in Table 3. The e¤ect of selling on FSBO is mostly not e¤ected and stays at roughly 4 percent. This is consistent with the numbers in Table 3 that suggested that while some characteristics were statistically di¤erent, the di¤erences seemed small. In column (iv) we also control for neighborhood characteristics by including neighborhood …xed e¤ects. The coe¢ cients on these controls are of no direct interest. However, the key is that we are able to explain 92.4 percent of the variation in the logarithm of price, and 89.3 percent of the variation in price. The impact of selling through FSBO goes down to approximately 3.2 percent, or 5,000 dollars. The regressions in columns (i) through (iv) focused on the impact of the channel through which the house was sold. In column (v) we explore the impact of the initial listing channel. There are two di¤erences relative to the results in column (iv). First, the sample now includes switchers: houses that initially listed on one channel but that sold through the other. These are mostly houses that listed on FSBO but ended up being sold through MLS. Second, now the FSBO dummy is de…ned as being initially listed on FSBO, as apposed to being sold 14

through FSBO. This regression is of interest for a potential seller asking what is the expected impact on price if they list on FSBO, and then behave optimally (depending on how lucky they were with the FSBO stock of buyers), regardless of where they end up selling. The results suggest that the premium for listing on FSBO, which is estimated at 3.1 percent, is almost identical to the premium for selling through FSBO. To further explore the distinction we also examine, in column (vi), the regression that includes both the initial listing channel and the sales channel. We see that there is a small additional – statistically signi…cant – premium of selling on FSBO of 0.7 percent. This premium is driven by the very small number of houses that initially listed on MLS, but were eventually sold on FSBO. In the last column we separate these houses and …nd that now the additional premium of selling on FSBO disappears, but that these houses command a large premium, over 6 percent relative to houses that listed and sold on MLS. Overall the results in Table 4 deliver a surprising result. Sellers on FSBO are able to sell their houses at a premium relative to MLS. In addition, sellers that initially list their houses on FSBO but that then move to MLS also command a signi…cant premium. The causal interpretation of the results relies on random assignment to platform, or random success, conditional on time, house and neighborhood characteristics. Random assignment is a strong assumption in this context. We deal with selection in the next section. Even after considering selection e¤ects we …nd that the commission is born by the sellers (see next section). We now examine other outcomes. In Table 5 we focus on the total time to sell. Time to sell is de…ned as the time between the initial listing and the sale date as recorded in the city data. The dependent variable in all regressions is the total time to sell, and the controls follow a similar structure to Table 4. In columns (i) through (iv) we focus on the sample of houses that sold on the channel where they were initially listed. Without any additional controls, the results in column (i) suggest that total time to sell is 4 days shorter when selling on FSBO. Once we control for year and month dummies, and for house and neighborhood characteristics, the e¤ect of selling on FSBO is not statistically signi…cant. The additional controls change the R-squared very little, compared to the sale 15

price where the house and neighborhood characteristics explained a large fraction of the variation.10 In the last three columns we once again study the full sample of houses that sold, not just houses that sold on the channel originally listed. In column (v) we …nd that sellers who originally list on FSBO should expect to take 20 days longer to sell. This is largely driven by houses that originally listed on FSBO but then switch to MLS. The results in column (vii) allow us to separate the e¤ects in four groups. The base group are properties listed and sold on MLS. Relative to this group the properties listed and sold on FSBO take 1 day longer, the same result we found in column (iv). For houses that listed on FSBO but eventually sold on MLS the time to sell is almost 64 days longer. This is not surprising since moving to MLS means starting from scratch. Finally, for houses that listed on MLS but that were sold through FSBO the expected time to sell is 120 days longer. To further characterize the di¤erences of outcomes between the two channels we report, in Table 6, the e¤ect of channel on the probability of sale. In all cases we regress a dummy variable, which varies by column, on channel dummy variables, year and month dummy variables, a linear time trend, house and neighborhood characteristics. We start by examining in columns (i) and (ii) the probability of a sale. The dependent variable is equal to one if the property sold. A non-sale is de…ned if we do not observe a sale price in the city data. Overall in the sample 87 percent of the properties sold. The properties initially listed on FSBO tend to have a higher probability of eventually being sold, although some of them are eventually sold through MLS. In column (ii) we separate the properties into four groups depending on initial listing and …nal channel. If the property sold the …nal channel is the channel where it sold, otherwise it is the last channel used for listing. We …nd that relative to the base group –properties that listed and sold on MLS – properties that sold listed and sold on FSBO are roughly 1.1 percentage points more likely to sell, although the di¤erence is not statistically signi…cant. The properties that listed on FSBO but eventually switched to MLS are even more likely to sell. Relative to the base group they are roughly 4 percentage points more likely to sell. The properties that list MLS 10

Time on market is de…ned by the timing of closing which depends on considerations hard to predict, thus a lower explanatory power is expected.


and switch to FSBO are less likely to sell, but this is an extremely small group and the e¤ect is not estimated precisely. In columns (iii)-(viii) we examine the probability of a sale, conditional on eventually being sold, within a …xed number of days. We look at 180, 90 and 60 days. We …nd a patterns similar to what we saw in Table 5: the properties listed on FSBO tend to take longer to sell. Thus, within a …xed interval of time a FSBO property is less likely to sell. Although FSBO listings are somewhat more likely to eventually sell, their initial success is lower than MLS. This is mainly driven by the properties that start on FSBO and switch to MLS. In columns (iv), (vi) and (viii) we separate the properties into four groups. The FSBO listing that sold on FSBO are less likely to sell within 60 or 90 days. This is consistent with MLS exposing sellers to a bigger stock of buyers (as in Claim 1). The properties that start on either FSBO or MLS, and then switch, take an even longer time to sell and thus are much less likely to sell within a …xed time period.



In the previous section we documented the di¤erence in outcomes for properties listed on FSBO and MLS. A key issue in interpreting the results is selection. There are two separate concerns. First, are properties sold on FSBO comparable to those sold on MLS? We control for a rich set of observed house characteristics, but it is still possible that there are unobserved di¤erences that are correlated with the platform choice. Second, even if the house unobserved characteristics are not correlated with the channel, the sellers attributes might be. We now discuss both of these issues in detail. 5.2.1

Unobserved House Characteristics

As we show in Table 2 there are some di¤erences in observed characteristics between the properties listed on FSBO and MLS. These di¤erences are not large but in some cases they are statistically signi…cant. Indeed, once we control for house and neighborhood characteristics, in the regressions we display in Tables 4-6, the results change somewhat. The di¤erences


in the observed characteristics might suggest that there are di¤erences in characteristics unobserved by us. To examine this issue we exploit properties that were sold multiple times in our sample using di¤erent platforms. As long as the unobserved characteristics are constant over time looking at properties that sold multiple times, and including a house …xed e¤ect will control for the unobserved characteristic In our sample, there are 2,020 properties that sold more than once. The majority, 1,869, sold twice, with 146 and 5 selling three and four times. Together this yields 4,196 sales. Out of these sales 3,371 (or 80%) were listed and sold on MLS, 628 (15%) listed and sold on FSBO, 194(5%) listed on FSBO and sold on MLS, and only 3 listed on MLS but sold on FSBO. Out of the 2,020 properties that were sold multiple times we have 645 that were sold using di¤erent channels in di¤erent times. In Table 7 we present results using this sample. Di¤erent columns focus on di¤erent outcome variables. In all regressions we include year and month dummy variables and a linear time trend. In almost all cases the results are similar to those we found in Tables 4-6, where we controlled for di¤erences across properties using the house and neighborhood characteristics. We also display in Table 7 similar regressions using the same sample, but dropping the …xed e¤ects and controlling for di¤erences using the house and neighborhood characteristics instead. Once again the results are essentially identical. Together these results suggest that there is no bias in the estimates due to an unobserved house e¤ect that is …xed over time. This should not be surprising. The di¤erences in the observed characteristics were not large and controlling for them did not make a large di¤erence. Since most unobserved house characteristics, that we can think of, seem (roughly) …xed over time we conclude that we should not be concerned over the impact of unobserved household characteristics on our estimates. 5.2.2

Seller Selection

If seller type a¤ects both price and platform choice our estimates will be biased. For example, some sellers might be better or more patient at bargaining and therefore able to get a higher price regardless of the channel they use. Being more patient, according to the model, they are also more like to list in FSBO. Absent appropriate controls for seller type we will overestimate 18

the e¤ect of selling on FSBO. We explore several ways to deal with this problem. The …rst approach is to compare the di¤erences in outcomes between those sellers who listed on FSBO and sold on FSBO and those that initially listed on FSBO but ended up switching to MLS. If FSBO listers are a more homogenous group than the sample as a whole, the selection problem should be attenuated; or eliminated if the success on FSBO is driven by factors unrelated to sellers’ type. The results in Table 4 suggest that conditional on listing on FSBO there is a small, and not statistically signi…cant, increase in price from also selling on FSBO. We …nd essentially the same result if we focus on the sample of initial FSBO listings. If we believe that moves to MLS, after listing on FSBO, are purely driven by random forces then the estimates suggest that the two platforms deliver the same prices. There is no gain in the sale price from selling on MLS relative to FSBO. Even if moving to MLS depends on seller type the selection bias should be reduced, as the group of FSBO listers is more homogenous than the population as a whole. Namely, in the range of sellers, these observations belong to the set that self selected into FSBO. Furthermore, it is not clear that the selection indeed dictates a bias. Consider selection on patience. Is it the more patient seller who moves to MLS or the less patient? A patient seller may stay longer on FSBO. On the other hand, moving to MLS entails a long wait (given the …ndings in the previous section), thus it might be that the more patient sellers are those that decide to move on to the MLS. In other words, there might be selection, but its relation to sales price is less clear. The results of time to sell and the probability of a sale, displayed in Tables 5-6, can not be directly compared for this purpose. Once a seller switches from FSBO to MLS it is only natural that it takes longer to sell. So it is not surprising that the total time to sell increases. Our second approach to quantify the role of unobservable seller characteristics is to compare FSBO sales to realtors’ transactions of their own properties. These transactions provide us with a "sale by owner" using the MLS. Levitt and Syverson (2006) report that realtors are able to obtain better prices when they sell properties in which they have an ownership stake relative to properties, sold by the same realtors, where they are not owners. 19

We assume that realtors are no worse at selling their own properties than non-agents. In other words, the e¤ect of realtors selling their own homes is an upper bound on the impact of seller selection. The results are presented in Table 8. The variable "Sold by Owner" is a dummy variable that equals one for all sales by either a realtor selling their own home on the MLS, or a sale on FSBO. The variable "Sold on FSBO" equals one for sales on FSBO, and therefore its coe¢ cient measures directly the di¤erence between the performance of FSBO sales and sales by owner/agents on MLS. The regressions in columns (i) and (iii) include only properties that sold on the channel where they were initially listed. The results in the other columns include all properties that sold. As in Levitt and Syverson we …nd that owners obtain a premium when selling properties in which they have an ownership share. However, for price, time to sell and probability of sale within 180 days there is no statistically signi…cant di¤erence between agent/owner and sales on FSBO. FSBO sales are less likely to happen within 60 or 90 days. Furthermore, if we control for heterogeneity across realtors in performance, we …nd that listing on MLS with top realtors ownership commands a premium over FSBO. This suggest that FSBO does not entail a penalty or premium and it is consistent with the …ndings from comparing FSBO listings only. We also examined instrumental variables regressions to control for the potential correlation between FSBO and the unobserved characteristics. In all these cases the impact of FSBO was not statistically di¤erent than zero. However, depending on the exact functional form the standard errors were very large, which is consistent with the instrumental variables being only weakly correlated with the decision to use FSBO. Indeed the "…rst stage" veri…es this. The instruments we tries include the neighbors’propensity to list, or their success, in FSBO. Our …nal approach is based on the fact that some sellers make multiple sales during the sample period. We use the observed multiple sales to control for unobserved seller heterogeneity. Matching names across transactions we identi…ed 265 sellers who listed properties using di¤erent channels, these involved 744 sales. The results are presented in Table 9. In the …rst column we regress the logarithm of price on a dummy variable that equals one if 20

the seller sold a property using FSBO any time during the sample, not necessarily at that observation. The sample includes all the sales in the sample and the regression includes the usual time, house and neighborhood controls. We see that most of the e¤ect of FSBO we saw in Table 4 can be explained by this dummy variable. This might not be too surprising since this coe¢ cient is a weighted average of the sellers that sold only once using FSBO and those that sold more than once and used FSBO at least once. Since the …rst group is larger they might explain most of the e¤ect. For that reason in column (ii) we run the same regression but constrain the sample to include only properties that listed on MLS. The results suggest that FSBO sellers are indeed likely to get a higher price even when selling through MLS. On average they get 1.6% more. Note, that they take on average roughly 10 more days to sell, suggesting that they are more patient. All this suggests that seller selection is indeed present. However, it is not enough to fully reverse the result that MLS does not command a premium. In the last two columns in the table we restrict the sample to the properties sold by sellers who had multiple sales/listings. In column (iii) we report the result of regressing the log of price, and time to sell, of the properties sold by these sellers on a dummy variable that equals one if the property was listed using FSBO, and the usual controls. We include also …xed e¤ects for the sellers. The results suggest that when listing on FSBO these sellers get 1.65% higher price, but the e¤ect is not statistically signi…cant. There is virtually no e¤ect on the time to sell. In column (iv) we repeat the analysis with a dummy variable that equals one if the property is listed and sold using FSBO. There are 225 di¤erent sellers that sold multiple properties using di¤erent channels involving 631 sales. As in column (iii) we include seller …xed e¤ects. The results suggest that there is no statistical di¤erence in the price and the time to sell is signi…cantly lower. In summary in this section we explored various ways to control for seller selection in the decision to use FSBO. The results suggest that indeed selection is present. After controlling for selection we …nd that the FSBO premium disappears, but there is no evidence that MLS provides any premium relative to FSBO.



Concluding Remarks

In this paper we examine the relative performance of two competing networks: MLS and FSBO. Our results suggest the two platforms command similar prices. Even after controlling for di¤erences in house and seller characteristics we …nd that MLS o¤ers no advantage in the sale price. This is not to say that using an agent is not valuable. Realtors can save sellers time and generally help through a stressful and maybe di¢ cult period. What do our results imply for market structure in the brokerage industry in Madison? If one believes that sellers are aware of the FSBO option, and know that there is no premium associated with MLS, then our results suggest that a large fraction of the population is willing to pay a signi…cant amount for the services provided by realtors. Thus, despite the 6% commission rate, realtors are going to continue to maintain a high market share. An alternative view is that is still di¤using. As more people become aware of it, and more importantly as more sellers realize that there might not be any price penalty associated with using it, its share of the market will increase. The data set we use in this paper comes from one market. We selected this market because of the availability of data and the willingness of and the local realtors association to cooperate with us and share their data. At this point we cannot generalize beyond this market. Without further data and analysis we do not know if our results hold more broadly. As we show the penetration rates of vary widely across neighborhoods. It is our impression, based on casual observation, that the penetration rates of FSBO vary across markets. Understanding what drives this variation and the forces behind the di¤usion of FSBO is key to understanding the broader implications of our …ndings. The data we analyzed so far end at 2004. We are in the process of cleaning data for 2005 and 2006. The importance of the additional data is that they allow us to study a market during a more di¢ cult time, during a cooler housing market. We could see if the cost or, returns to, using a realtor vary with the cyclicality of the market.


References [1] Coles, Melvyn and Abhinay Muthoo. “Strategic Bargaining and Competitive Bidding in a Dynamic Market Equilibrium" Review of Economic Studies 1998, 65, 235-60. [2] Coles, Melvyn and Eric Smith. “Marketplace and Matching" International Economic Review, 1998, Vol 39, No 1. [3] Follain, James, Terry Luter and David Meier. “Why Do Some Agents earn more than others" The Journal of Real Estate Research, Fall 1987, 73-81. [4] House, Christopher and Emre Ozdenoren. “Durable Goods and Conformity" mimeo University of Michigan. [5] Hsieh, Chang Tai and Enrico Moretti. “Can Free Entry be Ine¢ cient? Fixed Commissions and Social Waste in the Real Estate Industry" Journal of Political Economy 111(5), 2003 [6] Levitt, Steven and Chad Syverson. “Market Distortions When Agents are Better Informed: The Value of Information in Real Estate Transactions", University of Chicago mimeo. [7] Merlo, Antonio and François Ortalo-Magné. "Bargaining over Residential Real Estate: Evidence from England" Journal of Urban Economics, September 2004, 192-216. [8] Merlo, Antonio, François Ortalo-Magné and John Rust. "Models of Bargaining and Price Determination of Residential Real Estate: Theory and Evidence" 2006 mimeo University of Maryland. [9] Munneke, Henry and Abdullah Yavas. “Incentives and Performance in Real Estate Brokerage: Theory and Evidence" Journal of Real Estate Finance and Economics, 2001, 22, 5-21. [10] Salant, Stephen. “For Sale by Owner: When to Use a Broker and How to Price the House" The Journal of Real Estate Finance and Economics, 1991, 4, 157-173. 23

[11] Taylor, Curtis. “The Long Side of the Market and The Short End of the Stick: bargaining Power and Price Formation in Buyers’, Sellers’ and Balanced Markets" The Quarterly Journal of Economics, August 1995, 837-855. [12] Yavas, Abdullah and Peter Colwell. “Buyer Brokerage: Incentive and E¢ ciency Implications", Journal of Real Estate Finance and Economics, 1999, 18: 259-277


Table 1: Properties by Initial Listing Platform and Outcome, By Year ListnOutcome MLS FSBO Unsold Total 1998 to 2004 MLS 10,718 (85.9%) 31 (0.3%) 1,727 (13.8%) 12,476 (79.9%) FSBO 697 (22.2%) 2,153 (68.6%) 290 (9.2%) 3,140 (20.1%) Total 11,415 (73.1%) 2,184 (14.0%) 2,017 (12.9%) 15,616 1998 MLS 1,807 (84.2%) 3 (0.1%) 335 (15.6%) 2,145 (94.0%) FSBO 43 (31.2%) 77 (55.8%) 18 (13.0%) 138 (6.0%) Total 1,850 (81.0%) 80 (3.5%) 353 (15.5%) 2,283 2000 MLS 1,285 (87.0%) 4 (0.3%) 188 (12.7%) 1,477 (80.3%) FSBO 106 (29.3%) 226 (62.4%) 30 (8.3%) 362 (19.7%) Total 1,391 (75.6%) 230 (12.5%) 218 (11.9) 1,839 2002 MLS 1,456 (86.7%) 4 (0.2%) 220 (13.1%) 1,680 (76.8%) FSBO 99 (19.5%) 380 (74.7%) 30 (5.9%) 509 (23.3%) Total 1,555 (71.0%) 384 (17.5%) 250 (11.4%) 2,189 2004 MLS 1,564 (81.1%) 9 (0.5%) 355 (18.4%) 1,928 (74.4%) FSBO 103 (15.5%) 480 (72.4%) 80 (12.1%) 663 (25.6%) Total 1,667 (64.3%) 489 (18.9%) 435 (16.8) 2,591 The year is de…ned by initial listing date. An unsold property is de…ned as not having a sales price in the city data.


Area Area Area Area Area Area Area Area Area Area Area Area Area

Table 2: FSBO Penetration Rates, By Area FSBO Listing Share (%) FSBO Outcome Share(%) Properties Sold 70 43.6 33.7 101 28 39.3 25.0 56 17 37.7 28.1 231 89 33.8 27.0 148 19 27.7 18.1 155 1 26.5 18.7 219 21 24.5 16.1 143 2 20.9 14.1 206 88 19.4 12.6 325 76 17.8 12.9 309 39 12.2 7.7 181 73 9.4 7.3 382 86 7.9 2.4 165 Overall 20.1 14.0 15,808

An area is de…ned by the City of Madison for assessment purposes. The above areas are a sample out of areas de…ned by the city.


Table 3: Sample Property Characteristics by Listing Channel MLS FSBO Characteristic Mean Std. Dev. Mean Std. Dev. Di¤erence t-stat age (as of 2007) 46.35 24.39 48.29 26.46 1.94 3.71 # of bedrooms 3.07 0.72 3.05 0.68 -0.02 -1.46 # of full bath rooms 1.59 0.67 1.58 0.65 -0.01 -0.58 # of rooms 3.66 1.20 3.68 1.15 0.03 1.02 total sq footage 1,735.58 697.33 1,717.28 582.36 -18.30 -1.29 lot size 9,605.51 5,393.50 9,017.61 5,260.41 -587.90 -5.20 basement sq footage 998.43 381.88 959.41 329.76 -39.03 -4.98 inside condition 3.71 0.55 3.64 0.60 -0.07 -5.54 outside condition 3.75 0.49 3.75 0.51 -0.01 -0.90 roof age (as of 2007) 26.14 23.97 24.89 24.26 -1.25 -2.48 furnace age (as of 2007) 26.24 23.36 24.89 23.40 -1.35 -2.74 central air 0.81 0.39 0.82 0.38 0.01 1.63 quality class 4.79 1.16 4.84 1.07 0.05 1.90 street noise 16.13 26.90 15.35 26.47 -0.78 -1.38 water front 0.39 5.31 0.26 3.96 -0.13 -1.25 parcel view 2.03 0.20 2.02 0.18 -0.004 -0.94 The above characteristics are a sample of those available to us from the city data. Inside condition is de…nes as .., outside condition is de…nes as ..., quality class is... and parcel view is .... The sample include 12,476 houses listed with MLS and 3,140 houses listed with FSBO.


Table 4: The E¤ect of Channel on Price (i) (ii) (iii) (iv) (v) (vi) (vii) Dependent variable: logarithm of price Sold on FSBO/100 10.75 4.02 3.96 3.18 – 0.69 0.34 (0.92) (0.85) (0.35) (0.27) (0.45) (0.47) Initially Listed – – – – 3.05 2.55 2.82 on FSBO/100 (0.24) (0.40) (0.41) MLS Listing, – – – – – – 6.07 Sold on FSBO/100 (1.95) R2 = 0.015 0.190 0.870 0.924 0.925 0.925 0.927 Dependent variable: price (in 1000’s of dollars) Sold on FSBO 14.81 2.93 5.19 5.06 – 0.28 -0.57 (2.02) (1.93) (0.85) (0.71) (1.05) (1.08) Initially Listed – – – – 5.04 4.83 5.50 on FSBO (0.62) (1.05) (1.08) MLS Listing, 14.84 Sold on FSBO (5.07) R2 = 0.007 0.124 0.842 0.893 0.894 0.894 0.894 Time Controls no yes yes yes yes yes yes House Characteristics no no yes yes yes yes yes Neighborhood E¤ects no no no yes yes yes yes N = 12,871 12,871 12,871 12,871 13,599 13,599 13,599 All columns report results from OLS regressions. In columns (i)-(iv), the sample includes only properties that sold in the channel they originally listed. The sample in column (v) -(vii) also includes properties that sold on a di¤erent channel than originally listed. Time controls include year and month dummy variables and a linear time trend.


Table 5: The E¤ect of (i) (ii) Sold on FSBO -3.73 -3.65 (1.68) (1.67) Initially Listed – – on FSBO MLS Listing, Sold on FSBO Time Controls no yes House Characteristics no no Neighborhood E¤ects no no N = 12,870 12,870 R2 = 0.001 0.018

Channel on Time to Sell (iii) (iv) (v) (vi) -0.77 1.49 – -61.04 (1.63) (1.65) (2.83) – – 20.24 64.05 (1.51) (2.52)

(vii) -67.91 (2.91) 69.42 (2.57) 119.66 (12.10) yes yes yes yes yes yes yes yes yes yes no yes yes yes yes 12,870 12,870 13,598 13,598 13,598 0.148 0.173 0.179 0.207 0.213

All columns report results from OLS regressions. The dependent variable is total time to sell, measured in days, from the date of the initial listing until the sale date, recorded in the city date. In columns (i)-(iv), the sample includes only houses that sold in the channel they originally listed. The sample in column (v) -(vii) also includes houses that sold on a di¤erent channel than originally listed. Time controls include year and month dummy variables.


Table 6: The E¤ect of Channel on Probability of Sale (i) (ii) (iii) (iv) (v) (vi) Dependent variable: dummy Conditional on sale, sold variable equal to 1 if: Sold 180 days 90 days Initially Listed 1.90 – -7.58 – -11.75 – on FSBO/100 (0.69) (0.73) (1.10) FSBO listing – 1.09 – -0.49 – -2.52 sold on FSBO/100 (0.78) (0.81) (1.25) FSBO listing – 4.01 – -28.77 – -35.77 moved to MLS/100 (1.21) (1.23) (1.90) MLS listing – -3.90 – -20.94 – -7.93 moved to FSBO/100 (5.24) (5.65) (8.67) Mean of dependent variable(%) 87.1 87.5 53.0 N = 15,616 13,599 2 R = 0.134 0.134 0.132 0.161 0.118 0.134

(vii) (viii) within: 60 days -10.07 – (0.96) – -5.93 (1.09) – -20.94 (1.65) – -10.69 (7.56) 23.5 0.083

All columns report results from OLS regressions. The dependent variable is a dummy variable, which varies by column. In columns (i) and (ii), the sample includes properties that were not sold, while in columns (iii)-(viii) the sample is only properties that a sale was eventually observed. All regressions include year and month dummy variables, a linear time trend, house and neighborhood characteristics.



Table 7: House Fixed E¤ects Regressions Dependent variable: log of price time Initially Listed 2.44 2.63 – – 25.67 19.97 on FSBO/100* (0.49) (0.41) (3.47) (2.48) FSBO listing – – 2.34 2.71 – – sold on FSBO/100* (0.55) (0.46) FSBO listing – – 2.83 2.40 – – moved to MLS/100* (0.85) (0.75) MLS listing – – 4.23 3.44 – – moved to FSBO/100* (1.22) (3.87) House Fixed E¤ects yes no yes no yes no House+Neighborhood Char no yes no yes no yes

to sell –

Dependent variable: dummy Conditional on sale, sold variable equal to 1 if: Sold 90 days Initially Listed 0.30 0.52 -14.43 – – -7.38 on FSBO/100 (0.34) (0.24) (2.79) (2.51) FSBO listing – – – -6.01 -1.21 sold on FSBO/100 (3.15) (2.23) FSBO listing – – – -36.44 -37.63 moved to MLS/100 (4.87) (3.65) MLS listing – – – -31.96 33.68 moved to FSBO/100 (24.14) (18.75) House Fixed E¤ects yes no yes yes no yes House+Neighborhood Char no yes no no yes no

9.52 3.60 (3.88) (2.70) 68.82 70.89 (5.99) (4.42) 23.98 52.31 (45.08) (34.08) yes no no yes within: 60 days –

-2.48 -1.45 (2.84) (2.01) -20.33 -22.46 (4.38) (3.30) 6.77 -1.02 (21.75) (16.94) yes no no yes

*In columns where the dependent variable is "time to sell" the independent variables are not divided by 100. All columns report results from OLS regressions. The sample includes properties where multiple sales were observed, there are 2024 such properties involving 4206 sales. In columns where "sold" is the dependent variable the sample also includes properties that were listed more than once, at di¤erent times even if they did not sell, there are 2788 such properties involving 5030 listings. All regressions include year and month dummy variables and a linear time trend.


Table 8: FSBO versus Sales by Agent/Owner on MLS (i) (ii) (iii) (iv) (v) (vi) (vii) Dependent variable: log of price time to sell sold in 60 sold in 90 sold in 180 Sold by Owner/100 2.16 1.88 -1.10 -3.52 2.47 7.39 2.18 (0.70) (0.68) (4.19) (4.36) (2.75) (3.17) (2.08) Sold on FSBO/100 1.07 1.17 2.56 0.49 -6.82 -7.10 0.21 (0.74) (0.72) (4.42) (4.60) (2.90) (3.35) (2.20) N= 12,871 13,599 12,870 13,598 13,599 13,599 13,599 All columns report results from OLS regressions. In columns (i) and (iii), the sample includes only houses that sold in the channel they originally listed. The sample in columns (ii) and (iv) -(vii) also includes houses that sold on a di¤erent channel than originally listed. All regression include year and month dummy variables, a linear time trend, house and neighborhood characteristics.


Table 9: Controlling for Unobserved Seller Heterogeneity (i) (ii) (iii) (iv) Dependent variable: logarithm of price Initially Listed 1.65 on FSBO/100 (1.15) FSBO listing 1.38 sold on FSBO/100 (1.32) FSBO Seller/100 2.71 1.64 (0.25) (0.53) Dependent variable: time to sell Initially Listed 0.69 on FSBO (7.70) FSBO listing -21.68 sold on FSBO (8.84) FSBO Seller -0.82 10.51 (1.57) (3.25) Sample all sales MLS listings sellers w/ multiple listings/sales Fixed E¤ects no no yes yes N= 13,599 10,749 744 631 All columns report results from OLS regressions. In column (ii) the sample includes only properties there were listed on MLS. In columns (iii) and (iv) the samples include properties sold by sellers with multiple sales between 1998 and 2004, there are 265 sellers that sold properties listed using di¤erent channels, involving 744 sales, 225 seller sold properties using di¤erent channels, involving 631 sales. The regressions in columns (iii) and (iv) include seller …xed e¤ects. All regressions include year and month dummy variables. a linear time trend, house and neighborhood characteristics.