The middleman dilemma: exploring the relationship between formal ethics and ethics in practice among real estate agents in Sweden

Abstract

Purpose

The purpose of this paper is to explore the relationship between formal ethics and ethics in practice in the empirical context of real estate agents (REAs) working in the residential housing market, including owner-occupied houses and owner-occupied apartments, in Sweden. The paper investigates problems with the Swedish middleman model of real estate agency with regard to the acceptance among REAs of borderline professional behavior.

Design/methodology/approach

We report on a survey distributed to all Swedish licensed residential REAs to investigate their attitudes towards eight scenarios displaying borderline ethical behavior. Firstly, the means of each scenario were calculated, investigating signs of distance between formal ethics and ethics in practice. Secondly, logistic regressions were run for each scenario separately, thereby investigating factors affecting misconduct among REAs.

Findings

The empirical results show a clear difference between formal ethics and ethics in practice and also illustrate that some scenarios of borderline ethical behavior are creating greater problems for the REAs.

Practical implications

In Sweden, the seller is the principal, assigning the REA to sell a house or apartment, but the regulation is clear on the role of the licensed REA as responsible for promoting an informed and fair sales process where the buyer is safe to act without their own representative. Our study contributes with information to policymakers on possible areas for the development of the middleman model.

Originality/value

The paper is the first to empirically investigate the middleman model of a Swedish real estate agency in relation to the business ethics of the agents. The use of scenarios in close relation to the everyday working context of REAs as tests of ethics of practice is also of original methodological value to investigate possible diversions of professionals from national regulations.

Citation

Engström, R. and Söderberg, I.-L. (2024), "The middleman dilemma: exploring the relationship between formal ethics and ethics in practice among real estate agents in Sweden", Journal of European Real Estate Research, Vol. 17 No. 1, pp. 4-24. https://doi.org/10.1108/JERER-05-2023-0016

1. Introduction

For most consumers, the decision to sell or buy an owner-occupied house or apartment is seldom made, and thus, there is a need to use the services of local real estate agents (REAs). REAs work as mediators of residential markets all over the world. As rules and regulations differ across countries, the expectations of a REA’s contribution vary, as does what is perceived as sound business ethics within different market models. Formal ethics exists as formulated in national law and in recommendations by trade organizations, but research have detected signs of a distance between formal ethics and the ethics in practice. We therefore set out to investigate the existence and features of such a difference in order to get a deeper understanding of the nature of professional ethics within real estate agency. To also answer calls (e.g. Yavas, 2007; Zietz and Sirmans, 2011) for more knowledge on different national models for real estate agency, we specifically focus on the professional role of licensed REAs in Sweden, where the agents are hired by the seller but with a professional license to act as independent intermediaries in the sales process, so-called middlemen. This is regulated by law and the concept of generally accepted estate agent practice plays an important role. The buyer is in almost all cases not represented by an agent but is acting in own interest with sometimes limited knowledge. Trust in the professionality of the licensed REA is therefore crucial for this market to function without too much friction. To secure professional practice, the Swedish system for real estate agency relies on the Swedish Estate Agents Act (SFS, 2021, p. 516) and the authority of the Swedish Estate Agents Inspectorate’s (FMI) that issue licenses to the agents and also issues sanctions if needed.

The purpose of this paper is: (1) to investigate possible differences between formal ethics and ethics in practice in the context of Swedish real estate agency; and (2) to explore factors affecting this difference. The international use of two different concepts, broker and agent, is acknowledged, and for the purpose of this study, the word agent is chosen, as this is the concept used by regulators in Sweden.

The paper is structured in six sections. Relevant literature is presented in section two. In section three materials and methods for data collection and quantitative analysis are described and a basic model is presented based on relevant literature. In section four, results from the regression analyses are presented. Section five consists of a discussion on the implications of the results and, lastly, conclusions are presented in section six.

2. Literature review

2.1 Business ethics – formal ethics and ethics in practice

Business ethics can be seen as a concept that describes an approach to morality and ethics in the business world partly for an individual and partly for a company. The demand for ethics comes from different company stakeholders, such as customers and authorities, which also put pressure on company owners and board members to ensure a sound business. The value of good business ethics is difficult to measure. There are different views of business ethics, and the law is seen by many as the lowest level of business ethics. Ethical conflicts arise when there is uncertainty regarding the interpretation of regulations or where preferences differ across individual actors. There is a large body of research on different aspects of business ethics, and we only bring forward a few general aspects relevant to our research objectives and the focus on real estate agency.

Crane et al. (2019) highlighted the demarcation problem between ethics and law and presented a combination of the two in the form of a Venn diagram, where the two concepts overlap to some extent. They argue that the discussion on ethics should primarily focus on issues at the borderline of what could become law. The financial value of business ethics has been discussed among scholars, who have argued for a bonus for acting ethically. Paine (2000) questioned this approach and highlighted that ethics involves other aspects and merits a more thorough investigation than only focusing its financial benefits. The discussion on “greenwashing” in the last decade (e.g. Delmas and Burbano, 2011) is an example of this debate, with one argument being that if there is no bonus in communicating a green behavior, then companies will not spend their resources in keeping up a facade. Kaptein (2017) suggests that there is a constant struggle between ethics and the expectations and temptations in and around an organization. This struggle can take place within a person employed by the organization or between individuals and challenges the explicit ethics of the business.

LaFollette (2020) argued that all ethics must be put into practice in order to be effective and Clegg et al. (2007) developed a conceptualization of business ethics as practice, providing a practical application of ethics in organizations that goes beyond legal approaches. Velthouse and Kandogan (2007) empirically identified a large number of managerial skills and found that ethics was found less important among managers but in closer investigation was closely intertwined in everyday solutions of different practical tasks. This will serve as a short introduction to how the present study defines formal ethics and ethics in practice.

The formal ethics of real estate agency in Sweden will be covered in section 2.3. Here we review some of the literature on ethics in practice connected to real estate agency. Previous literature reporting on empirical studies on the ethics in practice or everyday business ethics among REAs is scarce (e.g. Stamsø, 2015; Brinkmann, 2009; Sirmans and Swicegood, 2000; Okoruwa and Thompson, 1999), and the main body of related research is conceptual and primarily focuses on market transparency (e.g. Fisher and Yavas, 2010; Kadiyali et al., 2014), sales incentives (e.g. Bernheim and Meer, 2013; Martin and Munneke, 2010; Bian et al., 2017; Levitt and Syverson, 2008), compliance with regulations (e.g. Levi and Terflinger, 1988; Nanda and Pancak, 2010) or misconduct (e.g. Pancak and Sirmans, 2008).

Sirmans and Swicegood (2000) found that female agents earn less than male agents and that gender related computer skills and access to an assistant had a significant impact on sales. Okoruwa and Thompson (1999) highlighted the effect of moral assessments impacted by education, income and gender, and Brinkmann (2009) described ethical problems in the everyday lives of REAs in a study on Norway. Nanda and Pancak (2010) pointed to the importance of interest groups, as their work often leads to a lower probability of state regulation; however, they also mentioned an insider and outsider problem, where licensed REAs may lobby for increased professional requirements and thereby reduce competition. Barker (2008) emphasized that increased educational requirements create barriers to establishment and points to high costs for requirements per year without consumer improvements. Levitt and Syverson (2008) illustrated agents’ conflicts of interest and examined how REAs act when selling their own homes, finding that they get paid 3.7% more than when selling the homes of others. The authors questioned why sellers do not conduct external independent valuations of their property. In contrast, Beck et al. (2018) found no evidence of a principal-agent problem.

2.2 Real estate agency

The most comprehensive review of the literature on real estate agency was conducted more than twenty years ago by Benjamin et al. (2000a, b) in an attempt to display different aspects of the phenomenon for further research. These works provided an annotated bibliography (2000a) and summarized the findings and the research front (2000b). The authors sorted the literature into six categories: nature of the market and of the firm; compensation issues, time on market and price, efficiency, liability issues and international market variations. They concluded (2000b) that there are differences in the legal structures surrounding real estate agency as well as understudied international differences calling for more research. Yavas (2007) pointed to the importance of real estate agency as a mediator in the transaction process and suggested the need for further studies using the lenses of contract theory, adverse selection theory and search theory. Zietz and Sirmans (2011) updated the earlier mentioned literature review with new findings, structured into six different categories: firm characteristics and franchising, commission and compensation, products and services, the effects on owner-occupied house and apartment prices and time to the market, technology and efficiency and productivity.

2.3 Real estate agency in Sweden

The specific market case—Sweden—is chosen because the role of the REA as an independent intermediary is rare from an international perspective and, therefore, interesting to describe and investigate further from the perspective of business ethics. There are a few other examples; Norway, that has a closely related system (e.g. Brinkmann, 2009; Lindqvist, 2008), and also, related in the dual role of the agent but not identical to the middleman, the model of dual agency offered in a few states in the United States. Kadiyali et al. (2014) provide an overview of the dual agency model and how it is regulated and used in practice in different states in the US. For further information on these different systems, we refer to Lindqvist (2008) who analyze transaction costs for single-family houses in an international comparison and illustrate regulations and market models in Sweden, Finland, Norway, Poland, England and the US. Jingryd (2013) compares the Latin market model to the agent model, and Delcoure and Miller (2002) compare real estate agency in the US with that in other countries. Engstrom et al. (2023) give a comparison of the professional development in Sweden and in Victoria, Australia.

The Swedish model of a middleman relies on the Swedish Estate Agents Act (SFS, 2021, p. 516), which stipulates the seller as the principal, with whom the agent draws a contract for his or her services and that the REA in taking on this role as coordinator in the market also has a duty towards possible buyers. As there are almost no buyer agents in the Swedish market they can be neglected when studying the functioning of the market.

To ensure the formal ethics, the FMI's supervision responsibilities include ensuring that REAs fulfill their obligations in accordance with the Swedish Estate Agents Act. The FMI includes a disciplinary board, which consists of members appointed by the government. Such supervision can lead to the decision that the REA should be issued a disciplinary reminder or a warning or even be revoked from the register. Ahlenius et al. (2021) gives a description of the educational level and working conditions of the REAs in Sweden.

To frame the empirical study, we briefly describe: (1) the formal ethics of the real estate agency model used in Sweden; (2) the home selling process; and (3) the Swedish real estate market of owner-occupied houses and apartments in which the respondents work.

First, we need to examine the regulatory framework. Supervision and legislation are extensive in Sweden to ensure the function of the market. A Swedish REA acts as an independent intermediary and with a duty of care (Estate Agency Act, chapter 3, section 1) against both buyers and sellers, but he or she is commissioned by the seller. Chapter 3, section 2 of the Estate Agents Act stipulates personal responsibility for the assignment agreement. The law (chapter 3, sections 3-4; 6–21) also states how a REA must act in reference to generally accepted estate agent practice. Ethical guidelines are formed through legal cases from the disciplinary committee at FMI (chapter 4, section 2 and 4). Some form of code of conduct can be found in most legal systems. Since the guidelines are shaped by legal cases (e.g. Melin, 2022), there may be some room for dynamism in the interpretation of sound estate agency practice.

The Estate Agents Act is binding for all registered residential REAs in the Swedish market. The law was enacted in 1984 and updated in 1995 based on academic studies (Melin, 2022), in 2011 (Jingryd, 2013; Melin, 2022) and most recently in 2021. A Swedish REA has advisory requirements linked to professional practice (chapter 4 section 12), where the agent can give advice to both buyers and sellers on matters that may be of interest to them. The agent also has an obligation to control (chapter 3, section 13) which property rights are linked to the home. This entails the ownership and what is included in the purchase. Ownership is controlled through a central property register, which is unusual from an international point of view. For condominiums, ownership is checked through a report from the condominium association. For properties and property extracts, there is information that is registered, as owner, size of the property, mortgages, communities, easement, assessed value, address and whether there is a decision on foreclosure, etc. If the REA does not check this, it will result in the penalty warning. If, for example, the ownership is not correct, then it becomes a question of a legal error, something that the REA can avoid through a far-reaching check.

In Sweden, the REA also engages in contract management (chapter 3 section 20), which in many countries is handled by a notary (Jingryd, 2013). The agent usually works with the exclusive rights of the contractual agreement for a maximum of three months, and hence, a consumer’s choice of which firm and individual will sell his or her home is significantly importance to the agent.

FMI is a central administrative agency under the Ministry of Finance, the tasks of which include registering, reviewing and supervising REAs and providing information regarding the concept of generally accepted estate agent practice. The supervisory activities of FMI can be thematic, authority initiated, or following from a public report. The public can file a report through FMI’s website.

Second, we need to provide a brief introduction to the practice of the sales process in the Swedish context, which begins with the agent’s getting an exclusive permit to mediate the selling of the home to a buyer. The parties then often plan showings of the house or apartment for sale. Most commonly, a bidding process then takes place, usually according to an English auction [1] (Rolli and Eberhart, 2005; Hungria-Gunnelin, 2018). The REA presents the bids to the seller, and the seller has free trial rights to choose the buyer (Jingryd, 2013; Melin, 2022). When the auction has ended and the seller accepts a preferred price and buyer, the REA draws up a contract for the sale [2]. The parties then go through the contract with appendices, like the object description, deposit agreement and extracts from the Swedish real estate register. The transaction most often ends with a physical so-called access meeting that is either carried out at the buyer’s lender (financial lending institute) or at the agent’s office, where the seller’s loan is redeemed, keys are exchanged and the application for legal registration when buying property is handled. This registration is only applicable if the transaction involves property, and thus, the buying and selling of condominiums does not entail such a registration. The REA must also discuss any return conditions, and the parties must be well informed about their rights and obligations, such as the buyers’ duty of inquiry and the sellers’ duty of disclosure. The buyer of a property usually performs an inspection; therefore, an inspection clause is an example of such a return clause. When the contract has gained legal force, a down payment is provided, and the buyer receives a bid list with the name, telephone number and time of bid to ensure the transparency of the bidding process.

Third, the Swedish market needs to be briefly described. In Sweden, the total turnaround market within the real estate agency industry dealing with private homes is 10.5 billion Swedish Krona [3] (Mäklarsamfundet, 2023), and approximately 92% of those who sell a single-family house choose to use a REA to carry out the selling of their home, despite the legal possibility of selling it themselves (Mäklarsamfundet, 2023). However, as Berggren et al. (2016) pointed out, private people are prevented from advertising with the leading housing advertiser, which might cause great problems for any seller wanting to sell a home without an agent (though this would not be the only reason).

There are approximately 7,400 REAs (FMI, 2024) registered in Sweden, but as a continuous monitoring of the number of active agents is lacking, data from non-official sources (booli.se, 2024) indicate that approximately 6,000 REAs are actively working in the Swedish market.

Presently in Sweden, a university degree of 2 years and an internship of 10 weeks are required as educational standards for those who want to become licensed by FMI. Internationally, there are large differences between the training requirements for REAs [4]; however, as this is not the main focus of the study, it is not covered in more detail here.

3. Method and data

This study is based on data collected in Sweden by a survey distributed in December 2018 and January 2019 by Kantar Sifo, an independent market research institute. The survey, comprising 56 questions in total, focused on individual background information, the experience of working as a REA in the home sales market, attitudes towards customers and the business and attitudes towards business ethics as illustrated by eight scenarios all displaying activities outside formal ethics. We selected all 7,085 registered (2018) REAs working as intermediaries in the housing market in Sweden. For a response rate of 20% of the whole population, 1,423 individuals completed the survey.

The data revealed the following demographic features: respondents’ ages ranged from 22 to 84 years, and the mean age was 44.5 years (standard deviation (SD) = 13.402). Out of the 1,423 agents, 43.2% were female and 56.7 male. The average number of years that respondents had been active as a licensed REA was 12.7 (SD = 10.157), with a minimum of 0 years for newly registered agents and a maximum of 54 years for those who have long been in the profession. 30.1% of the respondents were working in the Stockholm market (the capital of Sweden and the market with highest number of transactions per year as well as the highest prices per square-meter), while 69.9% worked elsewhere. The average number of sold condominiums and owner-occupied houses per licensed REA was 41.3 (SD = 27.517). A question on penalties for not complying to Swedish law indicated that 10.3% of the respondents had, at any occasion during their career, been sanctioned for this, while 89.7% had no earlier experiences of receiving penalties by FMI.

Inspired by the multidimensional scale (MES) first presented by Reidenbach and Robin (1988, 1990) and extended by Reidenbach et al. (1991), Reidenbach and Robin (1993), eight context-specific scenarios were constructed, copying the dilemmas in the original scenarios used by Reidenbach et al. (1991, p. 85) but constructed by the authors to capture the ethical problems inherent in the Swedish middleman model. We chose to construct scenarios that relate to actual situations being part of the idea of generally accepted estate agent practice, giving examples of when a consumer risk to be negatively affected by a REA’s ethical dilemma. The scenarios were carefully constructed based on three steps: (1) the cases were identified in the yearly reports from FMI’s disciplinary committee concerning their decisions regarding disciplinary actions against registered estate agents and agencies; (2) one of the authors have experience of 14 years as a licensed REA in Sweden and constructed the scenario descriptions; and (3) the scenarios were tested by a panel of four academics teaching within the REA program given at our department and four representative of real estate agency firms working as licensed REAs in Sweden. None of these representatives had any experience of being sanctioned by FMI and they all had more than 10 years of practice on the Swedish market.

The following scenarios, all considered as displaying a behavior that is outside formal ethics, where constructed (see Appendix for full-text context-specific scenarios):

1. The REA exaggerates the potential price to get the seller to sign a sales contract.

2. The REA misleads the seller to get repeated sales.

3. The REA hides information from the buyer.

4. The REA suggests to the seller a starting price level for advertising the property much below expected market value.

5. The REA misleads the seller in recommending a bank with unfavorable terms, so that the agent gets a personal kick-back.

6. The REA misleads the seller in recommending a bank with unfavorable terms, so that the real estate agency firm gets a kick-back.

7. The REA is recommending a specific property inspector to a client.

8. The REA uses a disclaimer clause as a standard in all contracts.

The eight context-specific scenarios were presented to the REAs in the survey with the instruction to rate each scenario on three different 7-point Likert scales with respective end points: unfair-fair, unjust-just and morally unacceptable-morally acceptable. One of the scenarios (scenario 5) was wrongly worded in the questionnaire and was therefore removed before analysis. Data were analyzed using IBM SPSS Statistics (version 27.0).

The three questions on each scenario were clustered together into one variable on moral equity per scenario. The seven remaining scenario variables had the following Cronbach’s alpha values: S1MOREQ = 0.918, S2MOREQ = 0.960, S3MOREQ = 0.983, S4MOREQ = 0.971, S6MOREQ = 0.976, S7MOREQ = 0.984 and S8MOREQ = 0.964.

Firstly, means of each scenario were calculated illustrating signs of a distance between formal ethics and ethics in practice. Secondly, logistic regressions were run for each scenario separately, to test the predicting power of six factors on this distance between formal ethics and ethics in practice. The following model was tested:

Rating of Scenario X=B0+B1 (GENDER)+B2 (AGE)+B3 (YEARS IN PROFESSION)+B4 (LOCAL MARKET)+B5 (NUMBER OF YEARLY TRANSACTIONS)+B6 (RECEIVED PENALTY)+e

4. Results

Removing the wrongly worded scenario, seven context-specific scenarios, all exposing behaviors that are not acceptable according to Swedish regulations and generally accepted estate agent practice, were introduced to the respondents. The REAs were all expected to answer regarding these scenarios as being unfair, unjust and morally unacceptable, that is, with very low ratings. The REAs’ ratings of the scenarios are shown in Table 1 (N = 1,423). We find that scenarios 1, 4, 6, 7 and 8 seem to represent cases where the Swedish regulations are not clear enough for respondents in comparison to scenarios 2 and 3, where the respondents seem to be aware of and/or respecting the regulations concerning what is an inappropriate behavior for a professional real estate middleman agent. As all of the displayed behaviors were outside of generally accepted estate agent practice the very lowest rating on the Likert scale could be expected. However, we decided to make 3 out of a 7-graded scale a breaking point as values above 3 definitely indicates that the respondent believes the activity to be commonly appearing within the profession, thereby illustrating a clear distance between formal ethics and ethics in practice.

As all of the scenarios are illustrating different aspects of generally accepted estate agent practice and thereby focus the studied difference between formal ethics and ethics in practice, we have run separate regressions for each of them. The results on factors affecting a distance between formal ethics and ethics in practice are here presented for each scenario and further discussed in section 5. As the same model of predicting factors have been used in all regressions only the first regression table (Table 2) is commented on in full in text, while the results of the regressions with the other scenarios as respective dependent variables are shown in tables (Tables 3–8) and only shortly commented in text. In the next section (section 6) these results are discussed.

Direct logistic regression was performed to assess the impact of factors on the likelihood that respondents would rate scenario 1 as acceptable (4–7 on the Likert scale), thereby expressing a clear distance between formal ethics and ethics in practice. The model contained six independent variables (GENDER, AGE, YEARS IN PROFESSION, MARKET, NUMBER OF YEARLY TRANSACTIONS and RECEIVED PENALTY). The full model containing all predictors was statistically significant, X₂ (6, N = 1,296) = 30.103, p < 0.001, indicating that the model was able to distinguish between those who reported a higher acceptance of the scenario and those who recognized it as unethical. The model as a whole only explained between 2.3% (Cox and Snell R2) and 3.1% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 63.8% of cases.

As shown in Table 2, only two of the independent variables make a statistically significant contribution to the model (AGE and LOCAL MARKET). The strongest predictor was LOCAL MARKET, recording an inverted odds ratio of 1.692. This indicates that respondents who were working on the Stockholm market were about 1.7 times less likely to report the nonethical scenario as just, fair and acceptable.

The full model of scenario 2 containing all predictors was statistically significant, X₂ (6, N = 1,377) = 19.628, p 0.003, explained between 1.4% (Cox and Snell R2) and 2.9% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 89.5% of cases. As shown in Table 3, only two of the independent variables make a statistically significant contribution to the model (GENDER and LOCAL MARKET), indicating that men and Stockholm agents were more inclined to accept this scenario.

The full model of scenario 3 containing all predictors was statistically significant, X₂ (6, N = 1,397) = 6.309, p 0.389, explained between 0.05% (Cox and Snell R2) and 0.08% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 63.8% of cases. As shown in Table 4, only one of the independent variables makes a statistically significant contribution to the model (NUMBER OF YEARLY TRANSACTIONS).

The full model of scenario 4 containing all predictors was statistically significant, X₂ (6, N = 1,334) = 126.757, p < 0.001, explained between 9.1% (Cox and Snell R2) and 12.1% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 62.5% of cases. As shown in Table 5, five of the independent variables make a statistically significant contribution to the model (GENDER, AGE, YEAR IN PROFESSION, LOCAL MARKET and RECEIVED PENALTY). Out of them being active on the Stockholm market, being man and having received a penalty from FMI predicted a higher likelihood of accepting the scenario.

The full model of scenario 6 containing all predictors was statistically significant, X₂ (6, N = 1,378) = 181.629, p < 0.001, explained between 12.3% (Cox and Snell R2) and 16.9% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 70.1% of cases. As shown in Table 6, three of the independent variables make a statistically significant contribution to the model (AGE, YEAR IN PROFESSION and NUMBER OF YEARLY TRANSACTIONS). The strongest predictor of considering the scenario as acceptable was having less years within the profession.

The full model of scenario 7 containing all predictors was statistically significant, X₂ (6, N = 1,391) = 137.625, p < 0.001, explained between 9.4% (Cox and Snell R2) and 12.6% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 63.7% of cases. As shown in Table 7, only two of the independent variables make a statistically significant contribution to the model (AGE and MARKET). Stockholm REAs were almost 1.8 times more likely to report the scenario as acceptable than REAs working on other markets in Sweden.

The full model of scenario 8 containing all predictors was statistically significant, X₂ (6, N = 1,360) = 57.728, p < 0.001, explained between 4.2% (Cox and Snell R2) and 5.8% (Nagelkerke R2) of the variance in the dependent variable and correctly classified 66.7% of cases. As shown in Table 8, only three of the independent variables make a statistically significant contribution to the model (GENDER, AGE and NUMBER OF YEARLY TRANSACTIONS). Gender was the strongest predictor and male REAs were 1.5 times more likely to find the scenario acceptable than female.

5. Discussion

The results showed that the REAs’ acceptance of the various violations of generally accepted estate agent practice described in different scenarios were higher than what would have been expected. The respondents are all licensed and can therefore be assumed to know the current regulations in Sweden (described in section 2, where the concept of “good estate agents practice” is clarified.). It is reasonable to assume that the REAs should consider all the scenarios to be unfair/incorrect/morally unacceptable (and therefore all rate 1). At the same time, there might be some variation in the understanding of the written text and the three different ethical aspects (justice; correctness and moral acceptance) that together created the level of acceptance can therefore maybe be perceived a bit differently between individuals so we decided not to be strict on reporting 1. Despite this, the REAs’ acceptance of the various violations of sound estate agents practice described in different scenarios is higher than what can be expected and we conclude there is a clear difference between formal ethics and ethics in practice.

The degree of acceptance is higher in some scenarios than in others. Only those scenarios where the respondents estimated an average value close to 3 are discussed here. This means that scenarios 2 and 3 described in the result section are not further addressed. Here, the significance of results regarding other tested scenarios (1; 4; 6; 7; and 8), which all are accepted to a greater degree indicating a greater difference between formal ethics and ethics in practice, are described and shortly discussed. As ethics in practice is context-specific we discuss the impact of the identified discrepancy between formal ethics and ethics in practice for Swedish consumers (sellers and buyers on the residential market).

• Scenario 1: The REA exaggerates potential price to get the seller to sign a sales contract.

A valuation of the owner-occupied house or apartment is part of the product offered to the customer by the REA. Marsh and Zumpano (1988) proposed relational aspects as reasons for REAs’ ability to use asymmetry. Agents create relationships with home sellers, bringing about trust and confidence in the related valuation and actions. Levitt and Syverson (2008) pointed out the conflict of interest in the valuation of sellers’ homes and argue that more sellers should hire an independent valuer with no interest in the transaction process, which is also relevant for the Swedish market. Ariely et al. (2003) show that consumers have difficulty valuing heterogeneous properties in homes (e.g. elevator, balcony and tiled stove or bathtub), which can lead to temptations for a REA to refer to data that is not relevant or no longer is current. Kaptein (2017) also points out that there are temptations at different levels and that such a temptation can be to overvalue an owner-occupied house or owner-occupied apartment in order to obtain a contract to sell. Nowadays, through good consumer access to market statistics via Internet, there is increased information about the final purchase price for houses and apartments, which facilitates the consumer’s opportunity to consider their own market valuation. Duarte et al. (2015) recommend the REA to try to contribute to increasing consumers' knowledge of the functioning of the housing market in order to avoid conflicts between the agent and the seller. One such conflict can be the view of the market value.

Our study shows that younger REAs have higher acceptance than the older for overvaluing a home in order to obtain an assignment from the seller. Hoyt et al. (2003) show similar results in their studies in connection with property valuations and suggest that this can be explained by the fact that longer experience and higher level of education reduces the probability of accepting this type of violation of the regulations. In a Swedish context, however, this is not a reasonable explanation. The effect of the number of years as a REA has been tested and the variable proves to be irrelevant as an explanation of the estimate. In Sweden, it is the younger agents who have the highest level of education as the education requirements have been gradually expanded (see section 2). The higher acceptance of initial overvaluation among younger REAs can instead possibly be explained by the fact that they are in the initial stage of building up their customer base to establish a personal brand.

• Scenario 4: The REA suggests to the seller a starting price level for advertising the property (owner-occupied house or apartment) much below expected market value.

The opposite of an overvaluation of a property is when a REA undervalues the owner-occupied house or apartment. While overvaluation is done to interest the seller in signing a contract on a sales assignment, undervaluation in advertising and other marketing of the object (so-called bait price) is a way to get more prospective buyers to the viewings and thus also get a faster and more intensive bidding. The study shows that the probability that a REA estimates this as acceptable increases if the agent is male; younger; have a shorter time in the profession; is active in the Stockholm area; and if the agent has previously received a disciplinary sanction from the FMI. The purpose of a bait price is to get a higher price to the seller and thus also a higher commission for the REA. Here, it is the buyer who is disadvantaged in that the participation in the bidding is based on an incorrect estimate of the market price that has been “rigged” by the REA to attract more bids. Using a bait price is contrary to the Swedish Marketing Act (2008, p. 486) and violates generally accepted estate agent practice.

International research has shown that a lower starting price can lead to a higher final price. Andersson et al. (2020) argue that a lower starting price leads to more bidders. More bidders would increase the information and reduce the risk of winners' curse by the bidder having information on how the rest of the market values the home. Milgrom and Weber (1982) also show that transparent bidding with public bids reduces the risk of winners' curse and thus also increases the willingness to pay. Kahneman and Tversky (1979) point out that new bids become new reference points that the bidder has to decide on. Whether a low starting price leads to a higher final price has been discussed in the literature.

• Scenario 6: The REA misleads the seller in recommending a bank with unfavorable terms, so that the agent firm gets a kick-back.

According to Swedish law, a REA must act carefully and objectively and take care of his or her client’s interests. Thus, recommending a bank that does not offer optimal terms is not in line with the law. The fact that the REA sells a service that is not advantageous to the consumer could be contrary to generally accepted estate agent practice and primarily the duty of care. In Sweden, the REA is allowed to mediate/sell ancillary services, while a number of countries, such as Norway, do not allow it at all (see section 2). Ancillary services can, for example, consist of mortgages, insurance, so-called “home styling,” moving services or advertising. The REA is permitted to receive compensation for this from the companies that are recommended, but the compensation should not exceed what is described as “a small value” and is regulated in Swedish law. What constitutes a small value may be assessed in each case. FMI believes that “the important thing is that the compensation is so low that it does not give any motive to take unfair consideration in the assignment” (Compare Government proposal 2010/11:15 p. 53).

The effect of the REA mediating ancillary services has been questioned in the Swedish market, as the effect may be increased difficulties for the consumer to value the core service (Berggren et al., 2016). Avery and Hendershott (2000) argues that such packaging of various services, so-called bundling, leads to price discrimination and Jud and Roulac (2001) suggest that sales of ancillary services (in the US) should be regulated. Unnecessary costs can arise for a consumer who trusts the REA and chooses to purchase these ancillary services, but it can also affect the trust of the profession, primarily with regard to duty of care and thus negatively affect the Swedish public’s image of the role as independent intermediary. Miceli et al. (2007) argued that various forms of conflicts of interest in information flow could be avoided if buyers and sellers each had their own representative.

• Scenario 7: The REA is recommending a specific property inspector to a client.

When a REA recommends a specific property inspector firm, it could therefore be misleading. There could be a conflict of interest if the cooperation between an agent and an inspector firm becomes so strong that the inspector feels pressure to adjust the inspection to obtain future contracts. Levi and Terflinger (1988) suggest increased transparency and less asymmetry in terms of requirements for examination and execution of house inspections. We find that younger REAs and REAs active in the Stockholm market estimate the agents’ influence on the choice of inspector as acceptable to a greater extent than older or non-Stockholm agents. If a collaboration or dependency relationship between a REA and an inspector is extensive, it could affect the requirement for objectivity. It can have extensive consequences for a buyer if the service is not exposed to competition.

• Scenario 8: The REA uses a disclaimer clause as a standard in all contracts.

This dilemma entails that the REA casually uses a disclaimer, which is permitted on special occasions, such as when moving abroad, or when a seller is placed under a trustee. To automatically place such a clause is contrary to generally accepted estate agent practice, as this majorly interferes with a buyer’s right to invoke hidden errors within the property. The buyer always has the right to claim against hidden errors when the seller must have had knowledge of the error.

Salzman and Zwinkels (2017) point out that consumers do not always act rationally or efficiently in connection with property transactions. The REA has a duty of care and a duty of information in order to reduce the asymmetry between the parties and give them the conditions to act more rationally in a stressful situation. A disclaimer can create extensive negative consequences for a buyer who cannot invoke hidden errors. Kurlat and Stroebel (2015) argue that the seller can also use an asymmetric information advantage about the condition of the home. A disclaimer clause leads to the market value of the home being negatively affected as the buyer’s option to invoke a hidden defect expires. It can also mean that a well-informed buyer can benefit from a price reduction as a result of a disclaimer, at the expense of a seller who is not familiar with the financial consequences that the clause may entail for the market value. A disclaimer can, of course, also mean future costs for the buyer that are impossible to discount.

6. Conclusions and directions for future studies

It is relevant to highlight the demarcation problem between ethics and law, as pointed out by Crane et al. (2019). They argued that ethics should primarily focus on issues at the border of what could become a law. This is in line with the concept of generally accepted estate agent practice used in the Swedish regulation. The supervisory authority is constantly monitoring the border between what is explicitly written by law and what is the practice of the community of national REAs. As times change, ideas in practice are changing, and the legal concept of “de lege lata” (the written law/what law is) has been discussed by experts to possibly be improved “de lege ferenda” (the wished-for law/what law should be), and minor or major changes have sometimes also been suggested based on these discussions to correct the written text.

We find it important to emphasize that a well-functioning national supervisory authority is crucial for preserving and, if necessary, increasing the status of the REA profession in any country, no matter the national market model. Although Sweden provides a context with a strong national supervisory authority and a clear regulatory framework, some flaws have been detected in the difference between formal ethics and ethics in practice of middleman REAs. Their attitudes towards some of the proposed activities, which are not in compliance with generally accepted estate agent practice, are outside of what can be regarded as acceptable by the national supervisory authority. This could be regarded as an alarm to national authorities and interest groups that a refined supervision of these specific issues is needed as a signal to professionals on the need to comply to formal ethics.

Different countries have different models of professional real estate agency, incentives in the market and regulations for consumer protection and competitive neutrality in the market, and there is a need for comparisons between these models to enhance market function and consumer protection globally. This study contributes with empirical evidence on some aspects of business ethics from one specific market model: the middleman REA in Sweden.

A small number of REAs in the Swedish market are disciplined with sanctions every year. This may not be alarming, as some flaws can be anticipated in all systems. Swedish REAs are cost efficient, as their transactions involve few parties to be paid with fees, but if agents have problems with their understanding of the concept of generally accepted estate agent practice it will affect customer trust and the residential market.

We also would like to point out the relational aspect of the profession studied, with its implications for trust and loyalty. More studies on these aspects are warranted, as are studies on the psychological and behavioral aspects behind consumer choices and the mediating role of REAs. To better understand the inclusion of information on mortgages and contacts with financial institutes or the different national models of incentives for agents also merits further study.

Table 1

Ratings by real estate agents/brokers of moral equity aspect of business ethics

Variable	Mean	S.D.	N
S1MOREQ	3.871	1.692	1,296
S2MOREQ	1.731	1.334	1,377
S3MOREQ	1.870	1.672	1,397
S4MOREQ	3.669	2.030	1,334
S6MOREQ	4.124	2.063	1,378
S7MOREQ	3.631	2.272	1,391
S8MOREQ	2.803	1.866	1,360

Source(s): Authors' own work

Table 2

Logistic regressions predicting factors affecting the likelihood of respondents reporting a distance between formal ethics and ethics in use for scenario 1 (the middleman real estate agent exaggerates potential price to get the seller to sign a sales contract)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.195	0.119	0.103	1.215	0.961	1.535
Age	−0.021	0.007	0.004***	0.979	0.966	0.993
Year in profession	0.013	0.009	0.149	1.014	0.995	0.993
Local market	−0.527	0.126	<0.001***	0.591	0.461	0.756
Number of yearly transactions	0.000	0.002	0.896	1.000	0.996	1.005
Receiving penalty	0.256	0.198	0.196	1.292	0.876	1.907
Constant	1.202	0.293	<0.001	3.328

Note(s): The dependent variable is SCEN1bin. The number of observations is 1,296. The Cox and Snell R² is 0.023; the Nagelkerke R² is 0.031. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 3

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 2 (the middleman real estate agent misleads seller to get repeated sales)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.428	0.188	0.023**	1.534	1.062	2.216
Age	0.002	0.011	0.852	1.002	0.981	1.023
Year in profession	−0.028	0.015	0.063	0.973	0.945	1.002
Local market	0.452	0.185	0.014**	1.572	1.094	2.259
Number of yearly transactions	−0.004	0.004	0.273	0.996	0.988	1.003
Received penalty	−0.052	0.318	0.871	0.950	0.509	1.772
Constant	−2.132	0.448	<0.001	0.119

Note(s): The dependent variable is SCEN2bin. The number of observations is 1,377. The Cox and Snell R² is 0.014; the Nagelkerke R² is 0.029. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 4

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 3 (the middleman real estate agent is hiding information from the buyer)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	−0.054	0.155	0.726	0.947	0.700	1.283
Age	−0.004	0.009	0.660	0.996	0.978	1.014
Year in profession	−0.006	0.012	0.614	0.994	0.970	1.018
Local market	−0.105	0.167	0.530	0.900	0.649	1.249
Number of yearly transactions	−0.007	0.003	0.045*	0.993	0.986	1.000
Receiving penalty	0.236	0.248	0.341	1.266	0.779	2.057
Constant	−1.150	0.379	0.002	0.317

Note(s): The dependent variable is SCEN3bin. The number of observations is 1,397. The Cox and Snell R² is 0.005; the Nagelkerke R² is 0.008. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 5

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 4 (the middleman real estate agent suggests to the seller a starting price level for advertising the property much below expected market value)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.418	0.120	<0.001***	1.519	1.200	1.922
Age	−0.015	0.007	0.032**	0.985	0.971	0.999
Year in profession	−0.021	0.009	0.024**	0.979	0.961	0.997
Local market	1.058	0.131	<0.001***	2.880	2.228	3.723
Number of yearly transactions	−0.002	0.002	0.270	0.998	0.993	1.002
Receiving penalty	0.490	0.198	0.013**	1.632	1.107	2.406
Constant	0.553	0.286	0.053	1.739

Note(s): The dependent variable is SCEN4bin. The number of observations is 1,334 The Cox and Snell R² is 0.091; the Nagelkerke R² is 0.121. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 6

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 6 (the middleman real estate agent misleads the seller in recommending a bank with unfavorable terms, so that the broker firm gets a kick-back)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.170	0.125	0.175	1.185	0.927	1.515
Age	−0.033	0.007	<0.001***	0.968	0.954	0.981
Year in profession	−0.036	0.009	<0.001***	0.965	0.947	0.983
Local market	0.081	0.134	0.549	1.084	0.833	1.410
Number of yearly transactions	0.008	0.003	0.003***	1.008	1.003	1.013
Receiving penalty	0.135	0.198	0.495	1.145	0.776	1.689
Constant	2.090	0.313	<0.001***	8.008

Note(s): The dependent variable is SCEN6bin. The number of observations is 1,378 The Cox and Snell R² is 0.123; the Nagelkerke R² is 0.169. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 7

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 7 (the middleman real estate agent is recommending a specific property inspector to a client)

	B	S.E.	Sig.	Odds ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.117	0.117	0.315	1.125	0.894	1.414
Age	−0.030	0.007	<0.001***	0.971	0.958	0.984
Year in profession	−0.026	0.009	0.005	0.974	0.957	0.992
Local market	0.563	0.126	<0.001***	1.756	1.372	2.248
Number of yearly transactions	−0.001	0.002	0.563	0.999	0.995	1.003
Receiving penalty	0.137	0.192	0.476	1.147	0.787	1.672
Constant	1.493	0.279	<0.001***	4.450

Note(s): The dependent variable is SCEN7bin. The number of observations is 1,391. The Cox and Snell R² is 0.094; the Nagelkerke R² is 0.126. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors' own work

Table 8

Logistic regressions predicting factors affecting the likelihood of respondents reporting a higher distance between formal ethics and ethics in use for scenario 8 (the middleman real estate agent uses a disclaimer clause as a standard in all contracts)

	B	S.E.	Sig.	Odds Ratio	95% C.I. for odds ratio
					Lower	Upper
Gender	0.426	0.123	<0.001***	1.532	1.204	1.949
Age	−0.035	0.008	<0.001***	0.966	0.951	0.980
Year in profession	0.012	0.010	0.231	1.012	0.992	1.032
Local market	0.201	0.128	0.117	1.222	0.951	1.571
Number of yearly transactions	−0.010	0.003	<0.001***	0.990	0.985	0.996
Receiving penalty	0.082	0.202	0.687	1.085	0.730	1.613
Constant	0.785	0.305	0.010	2.193

Note(s): The dependent variable is SCEN8bin. The number of observations is 1,360. The Cox and Snell R² is 0.042; the Nagelkerke R² is 0.058. *** = Statistically significant at the 0.01 level (2-tailed)

Source(s): Authors’ own work

Table A1

The context-specific scenarios used in the study

Scenario	Content	Full wording of the context-specific scenarios (all borderline to what is considered by regulators as good estate agent practice)
1	The middleman real estate agent exaggerates potential price to get the seller to sign a sales contract	A colleague of yours is very happy and tells that he/she has been on an intake and received the assignment to sell a property. The colleague had prepared carefully and had plenty of statistics and had also discussed housing prices in the specific area with colleagues in advance of the meeting. The seller’s focus was to find an enthusiastic real estate agent who really believed in the home and could get a high price. The colleague tells that he/she valued the home to SEK 6.5 million (app. USD 650,000), but the seller had told the colleague not to sell for less than SEK 7.5 million (app. USD 750,000) and the colleague had then tried to meet the seller’s wishes by adjusting the valuation to SEK 7.0 million (app. USD 700,000) in order to receive the assignment to sell the property
2	The middleman real estate agent misleads seller to get repeated sales	Your colleague has just – before the advertised viewing – conveyed a flat in his favorite property to a buyer who wants to buy, renovate and then sell the home for speculative purposes. The condominium had a clear need for renovation, but was located in the property’s best location with a view of the sea. The colleague tells us that the buyer did not want to bid the price the seller wanted, but that the colleague managed to talk the seller down to an acceptable level. The colleague will then be able to sell this condominium a second time within a six months period
3	The middleman real estate agent is hiding information from the buyer	Your colleague has just signed a clear contract between the seller and the buyer of a home in one of Stockholm’s more attractive suburbs. The seller made a pre-inspection, which means that an inspector examined the villa before sale. The bathroom was newly renovated by an established and large professional company. However, the careful inspector pointed out one construction error in the bathroom and remarked on this in an attached inspection report which was sent out both to real estate agent and to the seller. The seller refused to allow the surveyor’s inspection report to influence the sales process and pointed out that the purchasers, in fact, have a far-reaching investigation duty in Sweden and can now make their own inspection. The colleague thought it was a good proposal and all parties were now satisfied and happy at the contract writing (the buyers not aware of the inspection)
4	The middleman real estate agent suggests to the seller a starting price level for advertising the property much below expected market value	The colleague comes to the office and rejoices after having received an assignment in competition with five other real estate agents. The colleague, who is a really driven and enthusiastic real estate agent, wants to build a larger customer base, like many real estate agents. During the internship period, they have noticed that it is easier to build a customer base with many potential buyers on the display, since they are also prospective future sellers. The colleague’s market valuation of the home ended up at SEK 3 million (app. USD 300,000) and the starting price is set at SEK 2.7 million (app. USD 270,000) to get a good impression on the display
5 This question was excluded from the analysis due to a typo in the questionnaire	The middleman real estate agent misleads the seller in recommending a bank with unfavorable terms, so that the real estate agent, gets a personal kick-back	Your colleague conveys a home and succeeds in getting the buyer to choose the bank you cooperate with. According to a central agreement, the real estate agent at the agency firm receives a compensation from the bank that is percentage related to the size of the mediated mortgage loan. The amount that the colleague gets is 0.0005 of the purchase price (0.05%), which does not usually give any astronomical amounts but in this case it became good dividend. The current buyer borrowed SEK 6 million (app. USD 600,000), and this gives SEK 3000 (app. USD 300) in side income that the real estate agent, receives in his bank account
6	The middleman real estate agent misleads the seller in recommending a bank with unfavorable terms, so that the real estate agent firm gets a kick-back	The customer who is ignorant of bank loans now wants quick help in finding a lender. The customer does not receive any unique benefits through your bank cooperation. The colleague is curious about the list rates, looking at various banks' websites and noting that the list interest of your co-operative bank is just in the middle of the bank interest rates. For the colleague, it is obvious to always recommend your cooperation bank and follow the central agreement that exists between the real estate agency firm and the bank
7	The middleman real estate agent is recommending a specific property inspector to a client	You have had a long-term and well-functioning collaboration with an inspection company that also carries out your pre-inspections on properties in cases where this may be relevant. This inspection company also gives you a compensation for each individual mediated inspection of SEK 500 (app. USD 50). For your colleague, it is obvious to always recommend this property inspector
8	The middleman real estate agent uses a disclaimer clause as a standard in all contracts	Your colleague is an older and experienced real estate agent. He/she prefers to generally work with disclaimer clauses where the seller declines all responsibility for hidden defects in the property. He/she thinks that this facilitates the process and that it is nice to avoid future discussions. The sellers are part of this deal after the colleague has told them which terrible disputes can come in the future

Notes

1. An auction in which the auctioneer initially announces a low price and takes bids from potential purchasers until no higher bid is forthcoming (source: www.oxfordreference.com).

2. The bids are not binding in Sweden, unlike, for example, in Norway. There are no clear form requirements for bidding, in short, it can be said that the seller has all the rights and the buyer basically only has the right to make a bid. REAs must inform about how the bidding is done. This is often done in the prospectus that is handed out at the viewing, but it can also be found in the advertisements on the Internet. The seller receives ongoing information; the law is clear here. The buyer receives a bid list with name, phone number, bid, time and if there are conditions, they are called return conditions and can give one of the parties the right to request the return of the purchase given different conditions. There is no requirement for the REA to inform the other bidders if there is a bid that is dependent on a condition. It can be financing, inspection or selling your own home. This may mean that there are bids that the seller cannot accept, which affects the bidding process.

3. This equals approximately 11,339,000 USD.

4. To broker all types of housing for private use, 2 years of university and 10 weeks of internship are required, which from January 1, 2028, will be six months of internship and 3 years of university.

For commercial real estate, it is different. If the home is to be considered a significant part of a company, it may be mandatory to use a REA. The Estate Agents Act primarily concerns consumer protection.