NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Do Brokers Beat Auctions in the Sale of Foreclosed Real Estate Property? RCT Evidence from Sweden
Initial registration date
November 20, 2019
November 22, 2019 11:00 AM EST
This section is unavailable to the public. Use the button below
to request access to this information.
Other Primary Investigator(s)
Department of Statistics
The Swedish Enforcement Authority
Additional Trial Information
Extant research demonstrates that there is a foreclosure discount for real estates sold during foreclosure compared to similar real estates sold voluntarily at arms-length on the open market. In Sweden this involves a comparison of compulsory sale at a public auction administered by the Swedish Enforcement Authority (SEA;,swe. Kronofogden) with voluntary brokered sale. However, the impact of brokered sale during foreclosure has not been investigated previously. This is possible in the Swedish setting, since SEA is legally obliged to choose between these two different sale modes in each compulsory sale case.
Even though both modes are compulsory from the perspective of the foreclosed homeowner, the conditions for the sale are quite different. Auction sale is much formalized: the sale is advertised under SEA flag, there are limited opportunities for intending buyers to examine the property, and the auction takes place physically at SEA in an English-style auction, in most cases without a meaningful reserve price. By contrast, brokered sale mimics an open-market sale since the broker handles all aspects of the sale up until the signing of the contract. In both instances, the market value for each real estate is professionally appreciated by SEA. In this study, sale mode – auction or broker – is randomized by SEA for real estates that fulfil the legal criteria for brokered sale. Brokered sale is the treatment variable. The main outcome is sale price. This RCT design allows us to answer the following research questions: How do sale prices vary during compulsory sale of real estates on public auction versus brokered sale? How does market value impact sale price for these two sale modes? What are the implications of brokered compulsory sale for the foreclosure discount? The design also allow us to examine if failure to sale is different between the two sale modes.
Dahlberg, Matz, Mikael Lundholm and Mattias Nordin. 2019. "Do Brokers Beat Auctions in the Sale of Foreclosed Real Estate Property? RCT Evidence from Sweden." AEA RCT Registry. November 22.
Foreclosed real estate properties that will be sold under the authority of The Swedish Enforcement Authority (SEA), will be randomly assigned (by SEA) to one of two different sale modes; either they will be assigned to a broker to be sold on the private market, in an open-market sale (the treatment group) or they will be assigned to be sold in an English-style auction run by SEA (the control group).
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
1. Price: A continuous variable measured in tSEK available in all cases that results in the sale of the property.
2. Cost: A continuous variable measured in tSEK available in all cases. Costs include the cost for the professional market value appraisal, procedural costs, and brokerage fees. 3. Revocation: This is a dummy variable indicating whether the case is revoked by the creditors.
4. Sale: This is a dummy variable indicating whether the property is actually sold at auction or by a broker.
5. Duration: This is a continuous variable measuring the number of days from initiation of compulsory sale case handling at SEA to the sale of the property at auction or by a broker.
Primary Outcomes (explanation)
Regarding the cost variable:
a. The cost for the market value appraisal amounts to approximately tSEK 10 for a residential property and is charged in every case. Market value appraisers are publicly procured by SEA and the cost is predetermined by the contract.
b. The procedural costs are regulated in law and consists of a preparatory fee and a sales fee. The preparatory fee amounts to 1 per cent of the property’s tax value and is charged in every case at the point in time when either (i) the auction is publicly announced or (ii) SEA commissions brokered sale of the property. The sales fee amounts to 2 per cent of the tax value and is charged in a case that results in actual sale of the property in an auction. The minimum procedural costs amount to 20 % of price base amount (swe. prisbasbelopp), tSEK 9.3 per year 2019, and the maximum procedural costs amount to 150 % of the price base amount, about tSEK 70 per year 2019. The consequence of this regulation is that the preparatory fee in cases with brokered sale amounts to tSEK 9.3 for all properties with a tax value lower than tSEK 930. c. Brokerage fees are charged in every case that results in brokered sale of the property. These fees are determined according to a publicly procured contract and is a function of the sale price of the property. The current holder of the contract is Svensk Fastighetsförmedling, which is a brokerage firm with national coverage. The fee is set as a percentage of the sale price. The rate is the highest for properties with low values. As of year 2019 it starts out at 14 per cent of the sale price in the range of tSEK201-400 and is sequentially lowered to a 1.4 per cent if the sale price exceeds tSEK 5,000. If the sale price is tSEK 200 or lower, or tSEK 8,000 or higher, then the fee is individually negotiated for that specific property. There is a new public procurement planned for year 2020, which means that the brokerage fees may change during the duration of the study.
d. In summary, the costs are anticipated to vary between approximately tSEK 20 to tSEK 80 for properties sold on auction. For properties sold by a broker, the minimum costs are approximately tSEK 50. For brokered sale, there is no upper limit to the costs, as they depend on the sales price. As an example, the costs for a property with a tax value of tSEK 500 and sold by a broker for tSEK 700 amount to approximately tSEK 63. By comparison, the costs for a brokered sale of a property with a tax value of tSEK 2,000 and a sale price of tSEK 2,500 amount to approximately tSEK 84.
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
General Research Design
This is an RCT design with brokered sale as the treatment and auction sale as the control group. The sampling ratio to the treatment and control groups is 1:1. Inclusion criteria
SEA case-handlers determine if a specific property should be included in the study when the market value appraisal is delivered from the appraiser. The market value appraisal is a document describing the property and the appraised value. The appraiser always inspects the property physically if there are buildings on it.
The inclusion criteria are:
1. Property is legally eligible for brokered sale according to the Enforcement Code (ch. 12, par. 57)
2. Appraised value above tSEK 200
3. Property is registered as housing property or plot
4. No more than two registered owners.
5. No owner is deceased, if the estate is represented by more than two heirs.
6. There are no indications that the owner(s) will not cooperate.
These criteria are structured as a checklist for the case-handlers. This check list is annexed in Swedish (Annex A). The checklist enables the case-handlers to assess the inclusion criteria objectively. The case-handlers received training in brokered sale before the start of the study. They were then also informed about the study and the inclusion criteria by PI Lundholm.
SEA case-handlers perform the randomization immediately after determining whether the property should be included in the study according to the inclusion criteria. Randomization is carried out as follows:
1. The case-handler checks whether the last digit of the case identification number is odd or even. These numbers are assigned to each case in sequential order by the operative data program employed for registration and case handling immediately upon initial registration of the case. This initial registration is carried out by appointed case-handlers as their sole task, which means that they are not involved in the case handling at later stages. Accordingly, it is not possible for the case-handlers responsible for assessment of inclusion criteria and randomization to affect or change this case identification number. There are about 15 case-handlers in total involved in the randomization at the different geographic sections. 2. If the last digit is even the case is assigned to the treatment group. The case-handler registers this in the data program with a flag specifically designed for this study.
3. If the last digit is odd the case is assigned to the control group. The case-handler registers this in the data program with a flag specifically designed for this study.
The randomization protocol with instructions for the case-handlers in Swedish is found in Annex B. The case-handlers were informed about this by PI Lundholm at the same time as they were trained in brokered sale and informed about the inclusion criteria.
Experimental Design Details
Randomization is determined by the last digit of the case identification number (whether it is odd or even). These numbers are assigned to each case in sequential order by the operative data program employed for registration and case handling immediately upon initial registration of the case by appointed case-handlers. This is the sole task of these case-handlers and they are not involved in case handling at later stages. This means that it is not possible for the case-handlers responsible for assessment of inclusion criteria and randomization to affect or change this case identification number.
Real estate properties
Was the treatment clustered?
Sample size: planned number of clusters
Sample size: planned number of observations
Sample size (or number of clusters) by treatment arms
350 treatment properties and 350 control properties
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
See pre-analysis plan
Supporting Documents and Materials
There are documents in this trial unavailable to the public. Use the button below to request access to this information.
November 20, 2019
INSTITUTIONAL REVIEW BOARDS (IRBs)