How ESG directly and indirectly affects the financial performance of real estate companies?

(1)

How ESG directly and indirectly affects the financial

performance of real estate companies?

M.J.M. Meulenberg1

MSc. Finance University of Groningen Faculty of Economics & Business Supervisors: dr. A. Plantinga & dr. G. Alserda

Abstract:

In this study I examine the relation between Environmental, Social and Governance (ESG)- and financial performance for public listed real estate companies in the period 2009 until 2019. I use two analyses to study the dynamics of direct and indirect effects of the ESG-financial performance relation for real estate specifically. First, I asses whether best-in-class ESG portfolios are able to generate excess returns. Second, I propose operational performance as a mediator between ESG and financial performance. The results show that best-in-class ESG portfolios are unable to generate excess returns. I also document that best-in-class ESG portfolios are less sensitive to the market and engage more in aggressive investments. The second analysis documents significant mediation of operational performance. However, this effect is largely offset by the direct effect between ESG and financial performance. Because ESG scores are widely observable I conjecture that the effect of ESG on operational performance is already incorporated in the stock price.

JEL classification: G51, M14, D92

Keywords: ESG, real estate firms, financial performance, mediation and operational

performance

(2)

1. Introduction

In 2017, about 47% of all global listed real estate investment is owned by institutional investors. Of the global institutional assets base, around 50% is managed by integrating Environmental, Social and Governance (ESG) policies.2 Considering the scale of investments that integrate ESG policies, I study the financial performance of listed real estate companies by including real estate operating firms, real estate developing firms, real estate management firms and Real Estate Investment Trusts (=REITs) that have an active ESG strategy. To further advance our understanding of this relationship, I explore the effect ESG on financial performance through operational performance.

ESG are the three factors used to assess the sustainable and ethical impact of an investment in a business or company. Worldwide, investors are increasingly incorporating ESG policies in their investment decisions. A typical example of such an application could be a best-in-class investing strategy, by which investors only focus their attention on the top 25% or top 10% of the stocks in terms of ESG (Kempf & Osthoff, 2007). Empirical research regarding the effect of ESG on financial performance remains inconclusive (McWilliams & Siegel, 2002;

Cappelle-Blancard & Monjon, 2012). Most of this research investigates the effect of ESG on financial performance by using industry wide samples (Friede et al., 2015). Which is

surprising, given that this research finds significant differences across industries for the effect of ESG on financial performance (e.g., Waddock and Graves 1997; McWilliams and Siegel 2001; Fisman et al., 2005). According to Griffin & Mahon (1997) Multi-sector studies conceal industry specific effects. Subsequently, Chand (2006) suggests that studies analysing the link of ESG and financial performance should focus on a single industry. Therefore, I solely focus on analysing the relationship between ESG and financial performance for the real estate sector. ESG related investments are especially relevant in the real estate sector as real estate operations are responsible for 40% of global greenhouse gas emission. In addition to that, they are responsible for over 60% of the global usage of wood and adding to this3. As such, studies find that buildings with high efficiency ratings and better governance result in higher rents, higher premiums paid for the building, lower occupancy rates, and lower cost of capital (Eichholtz et. al, 2010; Fuerst & McAllister, 2011). Additionally, both ESG and the real estate sector focuses on long-term operations that may imply better results of ESG on

(3)

financial performance. However, current research either focuses on building level data or uses financial performance measures such as ROA and Tobin’s Q (Cajias et al., 2014; Eicholtz et al., 2013; Fuerst and Mcallister 2011). To extend the generalizability and increase the chance of promising findings (Friede, Busch & Bassen, 2015) I study the financial implications by focusing on stock performance which leaves the follow research question:

Do real estate companies with high ESG ratings generate excess returns?

However, most public investors are not fully informed about the ESG activities of a real estate company as this information is often not accessible, intangible and unstructured (Fiori et. Al, 2007). Therefore, Cai et. Al (2012) models the operational performance to assess the captured resources obtained by engaging in an active ESG strategy, for which they used cash flow from operating activities as a proxy for operational performance. In contrast with stock

performance, operating performance of real estate companies displays the interplay in cost and benefits for ESG. Analysing this interplay provides more insight in how the cost and benefit analysis of ESG on real estate is established. To date, no study has assessed the indirect relation of ESG on financial performance through operational performance in the real estate sector specifically. Therefore, to advance understanding the relationship between ESG and financial performance I do not only examine the direct relationship between financial performance and ESG but also investigate whether operational performance incorporates all or some effects that are eventually captured by financial performance. Accordingly, I propose the following research question:

Does operational performance mediate the relation of ESG and financial performance in the real estate sector?

To conduct this research, I construct a dataset of public listed real estate companies

worldwide over the period 2009-2019. For the first research question I assess monthly excess returns of real estate portfolios ranked on ESG scores obtained from ASSET4. The ESG scores are equally weighted and standardized which makes quantifiable analysis of portfolio returns possible. For this research question I construct best- and worst-in-class ESG

portfolios. I estimate the Fama & French 5 factor to adjust for risk which enables us to study whether alpha is generated (Fama & French, 2015).

(4)

ESG-financial performance relation. I use a structural equation modelling approach as proposed by the theory of mediation found by Baron and Kenny (1986).

The remainder of this paper is organized as follows, the next section covers the literature on the effect of ESG on financial performance, the relation ESG-financial performance for real estate specifically and the relation of operational performance and ESG. The third section describes the construct of ESG portfolios and model of the mediation of operational performance in the ESG-financial performance relation. The fourth section describes the ESG- and financial data. The final section describes the results and conclusion.

2. Literature Review

2.1 ESG and financial performance

In the early written literature, ESG had been labelled as a lead indicator for companies to generate deadweight losses and underperformance (Friedman, 1970). However, these early studies that focus on the question whether ESG could predict financial performance are based on weak theoretical foundations, inconsistent and inadequate measurement of ESG, weak methodology, and sampling problems (Ruf et al., 2001). Furthermore, Van Beurden and Gossling (2008) find that earlier studies on the relationship between ESG and financial performance incorporate too many interpretations of the period between 1970-1990, where ESG had low socio political value.

The incorporation of ESG, however, has become increasingly important for businesses over the past 2 decades. More than 50% of the global institutions now integrate ESG in their investment decisions (Kivits; Furneaux, 2013). This development clearly shows the salience of companies engaging in investments that are in accordance with ESG guidelines (Friede, Busch & Bassen, 2015). The integration of ESG policies in business models is also due to outside pressure. Eminent indices such as the S&P ESG indices, DJ Sustainability World Composite, or MSCI World ESG index monitor on the three different dimensions of ESG what encourages firms to integrate ESG in their business strategy. Ioana & Adriana (2013) find that around 80% of the 2200 largest corporations integrate ESG reporting in their annual report.

(5)

of environmental policies the more this increases the incentive to innovate. These innovations could ultimately lead to offsetting the costs of complying with environmental regulations. In addition to that, a strategy dedicated towards ESG could also result in greater loyalty of employees, customers, and local communities (Ribstein, 2005). Hence, lead to higher financial performance in terms of stock returns. Woods & Urwin (2010) also stress that the reputational aspect is one of the main reasons for integrating ESG in the business model. The reputational aspect of ESG is one of the main drivers for higher stock prices. The reputational value does not have any financial value on itself, but good reputation can be transformed into higher sales and eventually impact financial value. Otherwise the impact of ESG remains intangible. The reputational segment of investing is high on the agenda of institutional investors. Mostly pension funds engage in ESG related investments taking into account the social responsible investment policy that is in line with pension fund investments, hence, leading to a higher stock price for high ESG rated stocks (Lougee & Wallace, 2008).

2.2 Real estate and ESG

The real estate sector worldwide is accountable for 42% of the energy consumption (UNEP, 2016). There is an increasing demand for sustainability improvements by tenants, regulators, and investors in real estate. With respect to investors, this is especially the case for

institutional investors (Petersen & Vredenburg, 2009). Managing assets that are in conformity with an active ESG strategy is a priority on the agenda of institutional investors. Hence, it is important for the real estate industry to adapt, given the fact that the major shareholders in publicly traded real estate firms are institutional investors. As a result, the adaption in strategically managing ESG guidelines in real estate companies is increasing.

Pivo (2008) surveys almost 200 CEO’s of listed real estate companies. The result shows that the strongest drivers are: risk and return, outperformance and moral responsibility.

Interestingly, the literature assessing the link between ESG and financial performance for real estate companies is rather scarce. Nevertheless, several studies investigate the link for the real estate.

For instance, Cajias et al. (2011) find an increase in strength in ESG related activities among European listed real estate companies from 2000 till 2010, with REITS in particular.

(6)

governance and idiosyncratic risk. Furthermore, it triggers the information about the of expected cash flows.

McGrath (2013) studies the ESG and financial performance relationship for buildings

specifically. Results show that eco-certified properties have a lower excess capitalization rate than their non-eco counter properties. However, these results only find support by the energy star rating for buildings. For another rating, the so called LEED rating, results were much less conclusive. Both LEED certification and star rating are labels granted for energy efficient and sustainable buildings. The results for LEED ratings are not significant but do show a higher capitalization rate. This concludes that there are notable differences in pay-off between different sustainable certifications for buildings. In addition to that, it clearly shows the lack of consistency in results by not using standardized scores for ESG. Another study that uses LEED as an active ESG strategy proxy is by Sah et al. (2013). The authors differentiate between green REITs and non green REITs. The amount of buildings in the REIT portfolio that participate in the LEED program proxies the ESG activities and regarded as green REITs. They find that green REITs generate higher financial performance in terms of Tobin’s Q and ROA. Contrary to this finding, Mariana et al. (2018) report that the percentage of green certified buildings in a portfolio has a negative impact on the ROA, ROE and the alpha of stocks. According to the authors, this is due to incremental capital expenditures in order to receive LEED or energy star certification.

Because of these inconclusive findings, it is essential to further analyze whether ESG policies have an impact on the financial performance of a real estate company. The majority of

existing research focuses on property level and lacks control variables, which are likely to interfere with the reliability of the findings. Furthermore, most studies employ a sample of U.S. real estate companies and use different proxies for financial performance (Dermisi 2009; Eicholtz et al. 2013; Fuerst and Mcallister 2011; Riechardt et al. 2012).

(7)

2.3 Operational performance and ESG

Scholars find the effect of a positive relation between ESG and financial performance to be spuriously defined due to failure of identifying mediating effects of possible intangible resources (Surroca, et. Al, 2010). Various economists therefore stress the importance of operational performance, which entails the cash flows from operating activities, as it could provide the measurable missing link for the relation between ESG and financial performance (Caijas et al., 2011; Eicholtz et al., 2012).

First of all, a major aspect of making adjustments in firm structure towards a more active ESG strategy is the shift towards a longer term commitment in multiple areas, such as: stakeholder interests, sustainable development and the conditions regarding society. Studies establishing the long-term advantages of ESG investing find the positive effect on retaining employees (Suspanti et. Al, 2015). According to human capital theory, that suggests that when

employees would resign, a degree in knowledge embodied in labour productivity would be lost (Nafukho et al., 2004). Hence, this impacts operational performance of a company in multiple ways, such as loss of human capital and an increase in training costs for employees. Previous research on this subject matter indicates that ESG could increase the employee trust (Hansen et al., 2011). The increase in trust positively affects employee relationships, ESG therefore positively affects the social and human capital of the company. Eventually this process results in a significant increase of organizational effectiveness and efficiency that helps maintain the production capabilities at a high level (Holtom et al., 2008). Bauer et al. (2011) were among the first to study the relation between operational and environmental performance for listed real estate companies. Although there is a lack of causality, missing relevant control variables and a short time span, the relation still shows a positive and

significant effect. The effect, however, is only positive for real estate firms that operate in the office industry, whereas the residence sector performed relatively worse on environmental performance.

Another reason why ESG may impact financial performance through operational performance is by improving energy efficiency and the appliances installed in buildings that could

(8)

energy costs and, in turn, a higher operational performance. Therefore, I propose that the effect of ESG on financial performance is mediated by operational performance.

3. Methodology

To assess the effect of ESG on financial performance, I perform two types of analysis. First, I construct portfolios based on best-in-class and worst-in-class investing regarding ESG and asses compare their financial performance. Second, I examine whether operational

performance mediates the performance between ESG and financial performance following the mediation approach of Baron and Kenny (1986). Consequently, in the first analyse I analyse the direct effect and in the second analysis I determine the presence of an indirect effect.

3.1.1 Formation of ESG portfolios

Portfolio construction provides straightforward insights in how ESG scores in Real estate companies can be categorized (see also Wimmer, 2012). Hence, market value weighted portfolios are segmented into best-in-class and worst-in-class ESG scores. Thomson Reuters assigns ESG ratings at the end of each fiscal year. To be certain that either the ESG as well as the financial information is available, portfolios are constructed at beginning of July (𝑡) and hold for twelve month’s to the end June (𝑡$%_{) next year. The best- and worst-in-class}

portfolios contain a 25% best- and worst level filtered on the basis of ESG scores. These portfolios are named 𝐸𝑆𝐺_%) and 𝐸𝑆𝐺_*%) respectively. In addition to that, a 10% best- and worst in class portfolio is constructed named 𝐸𝑆𝐺_$+ and 𝐸𝑆𝐺_*$+. In addition to the 4 portfolios, the total sample of real estate companies is included, ESG100 respectively. The ESG100 acts as a benchmark comparison for the other portfolios. This leads to time series of monthly returns for the years 2009 till 2019. Depending on the applicable relative % level more real estate firms are added to the portfolios on an annual basis since more firms in the sample obtain ESG ratings later in the time span. The score of the threshold ESG determines the yearly rebalancing of the portfolios. Where the threshold is based on a relative %

(9)

3.1.2 Performance measurement of portfolios

The monthly excess returns of the portfolios are assessed using the Fama & French five Factor Model (Fama & French, 2015). It controls for the 5 different factors that cover size, value, profitability and investment patterns:

𝑅./− 𝑅1/= 𝛼.+ 𝛽+. 𝑅67− 𝑅17 + 𝛽$.𝑆𝑀𝐵/+ 𝛽%.𝐻𝑀𝐿/+ 𝛽<.𝑅𝑀𝑊/+ 𝛽>.𝐶𝑀𝐴/+ 𝜀./ (1)

where 𝑅_./ is the return on portfolio i in month t and 𝑅_1/ risk free rate from the Fama & French website at month t. The dependent variable in the model is the monthly portfolio returns 𝑅_./ in excess of the risk free rate 𝑅_1/. 𝑅₆₇is the monthly return of the MSCI world real estate index in month t. The HML covers the difference in return between a low and high book-to-market portfolio in month t. The SMB factor denotes the difference in return between small size and large size portfolios based on market capitalization in month t. The RMW factor denotes the difference between portfolios that are highly profitable minus portfolios that are weakly profitable in month t. At last the CMA factor denotes the difference between conservative and aggressive investment portfolios in month t. The purpose of this model is to specifically find outperformance that is adjusted for market risk and potential outperformance due to size, value, profitability and degree of investments. Consequently, this enables to study the link between the adjusted alpha and ESG.

Furthermore, according to Bauer et al. (2005) portfolios filtered on high ESG scores differ substantially from their conventional counterparts. Hence, it is relevant to asses any statistical differences in alpha between high and low rated ESG portfolios. Therefore, I test the excess returns on differenced portfolios (Derwall et al., 2005):

𝑅./C− 𝑅./*= 𝛼.+ 𝛽+. 𝑅67− 𝑅17 + 𝛽$.𝑆𝑀𝐵/+ 𝛽%.𝐻𝑀𝐿/+ 𝛽<.𝑅𝑀𝑊/+ 𝛽>.𝐶𝑀𝐴/+ 𝜀./ (2)

where 𝑅_./Cis the excess return of the high rated ESG portfolio and 𝑅_./* is the excess return of the low rated ESG portfolio. The independent variables in model (2) are similar to the ones in model (1). Except for the constant term 𝛼., that now captures the differenced alpha between

high and low rated portfolios. The differenced portfolio for the 25% and 10% class are ESG∆25

(10)

3.2 Robustness

3.2.1 Different benchmark

According to Löffler & Rapuach (2011) different systematic risk exposures can increase the generalizability of results. Therefore, as robustness a different market index, the ENGX is applied in the Fama & French 5 factor model. The ENGX exclusively contains REITS operating worldwide. However, Table I suggests a high correlation between the two market indexes and the regression results are therefore refrained and stored in the appendix.

Table I: Matrix of correlations

Notes: This table displays the correlation between the two market indexes over the period 2009-2019. Both the market indexes function to asses the market risk premium of the Fama & French 5 factor model.

3.2.2 Post financial crisis period

The results of the portfolio might lack consistency in terms of affection by the financial crisis in 2008. Therefore, market sentiment might bias the financial returns and ESG might not be valued correctly. According to D. Rommer (2014) global financial markets had affected equity markets until the end of year 2012. This would lead to inconsistency in results because of a potential structural break in the data. Therefore, a separate analysis for post financial crisis returns provides more conclusive results. The time period covers the period from the beginning of 2013 till 2019.

3.3 Mediating effects

In this section I examine whether ESG causes financial performance indirectly by operational performance. Figure I graphically displays the mediation analysis of this study. The graph shows operational performance as the mediating variable, financial performance as the dependent variable and ESG as the independent variable. Effect A multiplied by B is the indirect effect and effect C is the direct effect.

Variables (1) (2) (1) MSCI 1.000

(11)

Figure I: The relationship between ESG, financial performance and operational performance

Figure I: graphically shows the model for mediation analysis. The indirect effect is the product of effect A and B and the direct effect is effect C. Although, one must bear in mind that for ease of interpretation the control variables refrain from the figure. Only the independent variable, mediator and dependent variable are displayed.

Baron and Kenny (1986) state three necessary conditions that require for mediation to be present. First, the mediating variable must be significantly related to the independent variable. Second, the mediating variable must be significantly related to the dependent variable. Third, the dependent variable must be significantly related to the independent variable or diminishes the effect when the mediator variable is controlled for. However, the established effect between the independent and dependent variable in mediation might be misleading. According to (Zhao, 2010) this would represent the sum of the total effect, thus, including indirect and direct effects. Where in mediation the main focus is on measuring the indirect effect. Therefore, the regression of the independent variable on the dependent variable is refrained.

To establish a causal relationship with mediation one has to control for multiple variables that might lead to confounding the relationship between ESG performance and financial

performance (Baron & Kenny, 1986). First, firm size, since previous research indicates that larger firms receive more media attention and are therefore more prone to disclosing ESG reporting (Albers and Guenther, 2010). Moreover, larger firms are more likely to positively engage in shareholder activities (Dimson, Karakas & Li, 2015). Second, it is relevant to control the amount of leverage the firm has. The proxy is used to assess the risk tolerance that affects the priorities or tolerance towards ESG activities (Waddock & Grave, 1997). Third, an important variable to control for, especially for the valuation of real estate companies, is the amount of liquidity. Given the illiquid investments necessary for real estate investors a certain

(12)

threshold of liquidity is necessary to ensure pay offs to investors (Eichoholtz, 2001). Forth, dividend yield is included to assess possible investment opportunities (Wahba, 2010). Consequently, the following models are formed:

𝑂𝑃𝐸𝑅./ = 𝑎 + 𝛽$𝐸𝑆𝐺./*$+ 𝛽%𝑆𝐼𝑍./+ 𝛽<𝐿𝐸𝑉./+ 𝛽>𝐿𝐼𝑄./+ 𝛽)𝐷𝐼𝑉./+𝜀./ (3)

𝑅𝐸𝑇./= 𝑎 + 𝛽X𝑂𝑃𝐸𝑅./+ 𝛽Y𝐸𝑆𝐺./*$+ 𝛽Z𝑆𝐼𝑍./+ 𝛽[𝐿𝐸𝑉./+ 𝛽$+𝐿𝐼𝑄./+ 𝛽$$𝐷𝐼𝑉./+ 𝜀./ (4)

Where the subscripts t is the year and i the firm number. OPER refers to the operational performance measured by cash flows from operating activities divided by the total assets. ESG is the ESG rating from ASSET4. SIZ refers to the size of the company measured by taking the logarithm of the total assets. LIQ refers to liquidity and is measured by total cash divided by total assets (John, 1993). DIV is the dividend yield and RET is the yearly % change in the return index. At last 𝜀 captures the disturbance of the regression. For visual

interpretation of the metrics used in the mediation model see Table II.

I use structural equation modelling to make statistical interference in the model. Structural equation modelling uses a conceptual model, path diagram and system of linked regressions within observed and unobserved variables. SEM is fundamentally different from a normal regression where in normal regression there exists a clear distinction between dependent and independent variables. As opposed to SEM where dependent variables in one model can become independent variables. Furthermore, the SEM framework in mediation analysis is justified when extending the model to multiple independent variables (Gunzler, 2013). At last one of the assumptions for mediation by the Baron & Kenny approach is that the error terms in the different models need to be uncorrelated. As opposed to SEM where this assumption is less strict. Hence, this generates a higher power for mediation. A potential limitation in the regression between ESG and operational performance can be reverse causality. However, Giese (2019) finds that causality goes from ESG to firm performance. Therefore, exogeinity is more tenable in the model.

Table II: Description of variables used in mediation procedure

Name Computation

ESG-1 1 year lagged ESG score SIZ Logarithm of Total Assets

LEV Total Debt divided by Total Assets

LIQ Total Cash divided by total Assets RET Yearly % change of the return index DIV Yearly Dividend Yield

(13)

4. Data input/availability 4.1 ESG Ratings

The ESG data is obtained from the ASSET4 database. ASSET4 rates and collects ESG data on individual companies since 2002. More than 600 data points for each individual company, a labour intensive process that requires a single analyst a week for each individual company. All the data used to determine the ESG score is by screening exclusively objective and publically available data. For instance, the annual reports, stock exchange filings and NGO websites. The data is converted of key performance indicators and combined into eighteen category scores. Each of the eighteen levels is rated a score between 0 and 1 whereas a high score would indicate high performance on this measure and low vice versa. Ultimately, the 3 pillars environment, social and governance converge at an equally weighted basis and the integrated rating is obtained. The integrated equal weighted ESG scores range from 0-100 and follows the same screening as the 18 categories where a high number would indicate a high level of performance. Other data sources that report ESG scores such as Sustainalytics, GRESB (Global real estate sustainability benchmark) and MSCI ESG have been deliberated and analysed but were either lacking on reporting’s for listed real estate companies or covered a short time span. The major advantage of the usage of ASSET4 for research purposes is the computation of the equal weighing of all three pillars. This gives an unbiased standardized measure of the results when taking ESG as a whole. This standardized score allows for quantitative analysis of social responsible investments (Ribando & Bonne, 2010). The three different ESG components might have distinctive effects varying between different firms. For instance, the environmental component might be more relevant when analysing an

(14)

4.2 Financial and accounting data

Financial and accounting data are obtained from Datastream. The steps conducted to constitute the dataset necessary are as follows: I collected the ISIN’s of all the ESG dataset providers that covered real estate companies worldwide and pooled them together. ESG rating providers from MSCI ESG, Sustainalytics and GRESB. As such, I collected the ISIN’s of each individual firm summing up to around 3000 real estate firms. The real estate firms consist of REITs, real estate operating companies, real estate developing companies and real estate management firms. Real estate service firms are excluded from the dataset because real estate service firms do not generate income that is affiliated with real estate (Lindholm, 2006). This paper only studies listed firms because the research is oriented on stocks that are freely tradable. However more than half of the companies are not publically listed and therefore removed. The remaining number of firms are presented in Graph I. From the graph we see a steady increase of firms that receive ESG ratings. However, a drop of availability in 2018 indicates that some firms had not received ESG ratings yet by the beginning of June 2018.

Graph I: Total amount of companies with ESG rating

Notes: this table shows the total sample of real estate companies that retrieved an ESG rating from 2009-2018. Note that the ESG25, ESG-25, ESG10 and ESG-10 are constructed based on this sample.

As already mentioned the objective is to collect a sample that covers real estate companies worldwide. However, most companies that are operating in the real estate business and denoted on a stock exchange are operating in developed countries. The distribution of the countries covered is visible in Graph II.

(15)

Graph II: Distribution of countries for the covered real estate companies

Notes: graph 1 displays the distribution of countries of which the total sample of real estate stocks is made of. The whole period of 2009 till 2019 is covered in this distribution. Countries with less than 0.5% share are added to the rest of the world section.

The financial and accounting obtained from Datastream includes data on market

capitalization, total assets, cash flow from operating activities, return indexes, total debt, total cash and dividend yields. The Fama & French factors are downloaded from the Fama French website. The monthly return indexes and Fama & French factors are downloaded from the beginning of each month. The factors of Fama & French are the 5 factor developed index which is justified as Graph II indicates the high frequency of developed countries.

4.3 Descriptive statistics

Table III presents the summary statistics of the monthly excess portfolio returns. Panel A displays the whole time span (2009-2019) and panel B covers the post-financial crisis period (2013-2019). The mean excess returns of the high rated ESG portfolios are in both filters higher than the low rated counterparts. The table shows positive returns for all portfolios with returns varying between 0.80% and 0.95%. The returns in the post financial crisis are lower compared to the returns of the whole period which indicates the negative impact of market sentiment. Moreover, the ESG100 portfolio generates the lowest returns indicating a U-shaped relation between stock returns and ESG which is in line with the finding of Barnet & Salomon (2012). They argue that the ability to profit from ESG depends on stakeholder influence capacity (SIC). Accordingly, firms with low ESG performance generate negative returns as they invest more in ESG until the the relationship neutralizes and turns positive as SIC accrues and triggers better financial performance stemming from higher ESG expenditures.

Brazil Singapore Cayman Islands Japan Hong Kong

(16)

Table III: Summary statistics of monthly portfolio returns (%)

Variable Mean Std.Dev. Min Max

Panel A: 2009-2019 SMB _-.028 _1.331 _-3.310 _3.950 HML -.134 1.773 -4.450 4.700 RMW .305 1.133 -2.800 3.330 CMA .004 1.043 -2.710 2.730 Rm-Rf .958 4.125 -11.398 12.167 ESG100 .764 5.938 -13.406 17.496 ESG25 .878 4.797 -14.406 13.252 ESG-25 .764 6.839 -13.283 18.906 ESG10 .949 4.649 -16.129 14.012 ESG-10 .810 7.923 -24.843 21.015 ESGΔ25 .114 5.820 -19.339 14.398 ESGΔ10 .139 7.299 -16.240 24.477 Panel B: 2013-2019 SMB _-.093 _1.305 _-3.310 _2.760 HML -.146 1.753 -4.450 4.350 RMW .204 1.043 -2.800 2.790 CMA -.072 1.052 -2.620 2.730 Rm-Rf .704 3.427 -8.687 10.340 ESG100 .878 6.072 -13.406 17.496 ESG25 1.099 4.115 -7.553 13.252 ESG-25 .956 7.261 -13.283 18.906 ESG10 1.232 3.469 -7.201 8.673 ESG-10 .969 8.397 -24.843 19.450 ESGΔ25 .143 6.641 -19.339 14.398 ESGΔ10 .264 7.665 -16.027 24.477

Notes: Table shows the percantage of monthly excess returns. Panel A displays the portfolios for the whole covered time span

2009-2019. Panel B covers the portfolios for the time span of 2013-2019. Where Rm-Rf displays the MSCI World Real estate index return in excess of the risk free rate. ESG100 covers the whole sample of firms that received an ESG rating. ESG25 and ESG-25 covers the 25 best and 25 worst-in-class respectively. ESG10 and ESG-10 cover the 10% best- and worst-in-class respectively. ESGΔ25 and ESGΔ10 cover the differenced portfolios for the 25% and 10% respectively.

Table IV displays the descriptive statistics of the variables applied in the mediation analysis. However, one must take notion that these descriptive statistics are part of a longitudinal dataset with yearly changes instead of monthly changes. This enables us to add more control variables in the model since most data available in Datatstream only mutates yearly. The table displays the total observations (N), the mean, standard deviation, minimum and the

(17)

Table IV: Descriptive Statistics of mediation panel data

Variable N Mean Std.Dev. Min Max

ESG 4029 39.65 28.83 2.82 96.28 SIZ(log) 3762 16.28 2.11 11.29 24.76 LEV(%) 3761 38.50 21.40 0 537 LIQ(%) 3598 5 7.02 0 89.60 RET(%) 3671 9.95 29.53 -98.66 548.95 DIV(%) 4529 4.20 3.07 0 36 OPER(%) 3760 3.09 5.10 -.42 .47

Notes: The table displays the different set of variables used in the the panel dataset with the applicable descriptive statistics.

This set is exclusively used for the mediation analysis and denotes me. Where ESG is the one year lagged ESG score, SIZ the firm size, LEV the amount of leverage, LIQ the liquidity, RET the return index, DIV the dividend yield and OPER the operational performance. The time span covers 2009 till 2019 and the variables cover yearly results for individual firms.

5. RESULTS

5.1 ESG Portfolios

Graph III presents the portfolio returns for the ten-year period from July 2009 till July 2019. The monthly portfolio returns in excess of the risk free rate are indexed at the beginning of the period to display the compounded returns. The graph shows higher returns for the best-in-class portfolios compared to the worst-in-best-in-class portfolios. It also shows that the ESG25 has higher returns than the ESG10 portfolio. This indicates that higher rated ESG portfolios are more rewarding.

Graph III: Indexed returns of different portfolios from 2009-2019

Notes: This table plots the excess returns generated by the 4 high and low ESG portfolios filtered on a 25% and 10% cut off, the excess return of the portfolio that contains the total sample size of real estate firms and the excess return of the MSCI Real estate index which is used as the market return. The returns are index at the first day of July in 2009 and covers up the returns until end of June in 2019. Table V summarizes the results for the Fama & French 5 factor model estimation for the total ESG rated sample (ESG100), the best-in-class ESG portfolios for the 25% and 10% (ESG25 and ESG10 respectively) the the worst-in-class ESG portfolios for the 25% and 10% (ESG-25 and

0 50 100 150 200 250 300 2009 2010 2011 2012 2012 2013 2014 2015 2016 2017 2017 2018

(18)

ESG-10 respectively) and the differenced portfolios for the 25% and 10% (ESG∆25 andESG∆10 respectively). I control for heteroskedasticity by using white standard errors. I test for the presence of serial correlation by performing the Breusch Godfrey test. This test shows no presence of serial correlation. The results show that the market risk, operating profitability factor, the book-to-market and the investment factor all have a significant impact on the portfolios. All the portfolios show a positive alpha, however, only the 25% best-in-class portfolio is positively different from zero4. In addition to that, the alpha only shows a

marginal positive percentage. Consequently, the result in both differenced portfolios show no significant difference between best- and worst-in-class portfolios. Hence, I fail to find

evidence of a direct association between ESG and higher financial performance. One has to bear in mind that the correction for the 5 factors is largely responsible for this insignificant result. However, the factor loadings can still provide valuable insights for real estate

companies. The market risk is significant for all the constructed portfolios, where high rated ESG portfolios have lower market risk than low rated portfolios. The differenced portfolios, confirm the significant difference at a 5 and 10 % level in market risk between the high and low rated ESG portfolios. This finding is in line with multiple studies, which show that firms with a high degree of responsibility towards ESG are less exposed to overly negative market reactions or regulatory actions against them (Godfrey et al., 2009; Sassen et. Al, 2016; Eichholtz, Kok & Yonder, 2012). As for the HML coefficient that concedes negative factor loadings for both worst-in-class portfolios. This means the low rated ESG portfolios hold more growth stocks in their portfolio. This study is in line with current research that suggests that mature firms are more likely to incorporate ESG polices compared to growth firms (Tamimi & Sebastianelli, 2017). The differenced portfolios confirm this finding with 10 % significance in both cases. Furthermore, the results show negative factor loadings on the profitability factor coefficient RMW. Therefore, the sample of portfolios consists of

companies with weak operating profits. At last, the CMA coefficient is negative for the low rated ESG portfolios. Both differenced portfolios show a statistical significance at a 5 % level, this indicates that firms which integrate ESG policies pursue more aggressive investments. This finding is in line with Mariana et al. (2018) who find that integrating ESG increases capital expenditures significantly.

(19)

Table V: Performance level portfolios in a multifactor regression model, 2009-2019

Coefficient 𝐸𝑆𝐺_$++ 𝐸𝑆𝐺_%) 𝐸𝑆𝐺_*%) 𝐸𝑆𝐺_$+ 𝐸𝑆𝐺_*$+ ESG∆25 ESG∆10

α 0.0861 0.694* 0.003 0.465 0.203 0.691 0.262 (0.444 (0.372) (0.539) (0.291) (0.607) (0.507) (0.649) ß0 (ß) 0.873*** 0.593*** 0.888*** 0.737*** 0.870*** -0.295** -0.132* (0.101) (0.112) (0.115) (0.0947) (0.137) (0.113) (0.090) ß1 (SMB) -0.536 -0.333 -0.454 -0.0959 -0.677 0.121 0.581 (0.324) (0.293) (0.400) (0.260) (0.488) (0.363) (0.521) ß2 (HML) -0.810** 0.113 -1.116** -0.102 -1.389** 1.229** 1.287** (0.407) (0.412) (0.491) (0.364) (0.550) (0.503) (0.569) ß3 (RMW) -0.938* -1.238** -0.842 -0.778* -1.429* -0.396 0.651 (0.539) (0.501) (0.649) (0.439) (0.801) (0.629) (0.806) ß4 (CMA) 1.032** -0.0637 1.368*** -0.247 1.396** -1.432** -1.643** (0.427) (0.502) (0.517) (0.397) (0.656) (0.687) (0.707) Observations 120 120 120 120 120 120 120 R-squared 0.428 0.396 0.338 0.496 0.274 0.149 0.064

Notes: This table summarizes for each portfolio the factor coefficients, the R-squared and the alpha’s using the Fama French 5 factor model. The coefficients ß0, ß1, ß2, ß3, ß4 refer to the factors reported in the brackets next to the coefficient. The

portfolios are value-weighted. The ESG100 portfolio consists of all ESG rated real estate companies in the dataset. The ESG25

and ESG-25 consist of the highest and lowest rated 25 % ESG rated real estate firms in the dataset. The ESG10 and ESG-10 are

the highest and lowest rated ESG rated real estate firms based on a 10 % level. The ESG∆25 is the differenced portfolio the 25

% ESG robust standard errors are between brackets reported under the applicable coefficients. The time span is based on 2009-2019. ***, **, * indicate significance levels based on 1, 5 and 10 % level.

5.2.2 Results for post financial crisis period

During the financial crisis the ESG portfolios might have been influenced by market

(20)

Table VI: Performance level portfolios in a multifactor regression model, 2013-2019

α 0.373 0.918** 0.400 0.724*** 0.430 0.518 0.294 (0.520) (0.355) (0.653) (0.269) (0.742) (0.667) (0.760) ß0 (ß) 0.991*** 0.703*** 1.028*** 0.745*** 1.101*** -0.325* -0.355* (0.147) (0.109) (0.178) (0.0595) (0.205) (0.177) (0.213) ß1 (SMB) -0.837* -0.533* -0.911 -0.113 -1.098 0.378 0.985 (0.447) (0.281) (0.562) (0.238) (0.675) (0.566) (0.673) ß2 (HML) -1.376** -0.125 -1.751** 0.392* -2.018** 1.625* 2.410*** (0.671) (0.343) (0.873) (0.233) (0.989) (0.855) (0.904) ß3 (RMW) -1.638 -1.776*** -1.678 -0.0534 -2.332 -0.0980 2.278 (0.991) (0.508) (1.286) (0.383) (1.597) (1.283) (1.469) ß4 (CMA) 1.900*** 0.268 2.306*** -0.577* 2.192** -2.038** -2.768*** (0.534) (0.611) (0.713) (0.321) (0.881) (0.984) (0.851) Observations 84 84 84 84 84 84 84 R-squared 0.409 0.431 0.339 0.541 0.293 0.175 0.155

and ESG-25 consist of the highest and lowest rated 25%t ESG rated real estate firms in the dataset. The ESG10 and ESG-10 are

the highest and lowest rated ESG rated real estate firms based on a 10 % level. The ESG∆25 is the differenced portfolio the

(21)

5.3 Mediation of operational performance

The portfolio results show that it is not feasible to generate excess returns directly by investing in a best-in-class ESG portfolio. Therefore, no positive direct effect between ESG and financial performance is found. Table VII presents the results of the total effect of ESG on financial performance after controlling for dividend, size, liquidity and leverage. The graph shows an insignificant positive ESG coefficient, which is in line with the results of the best-in-class ESG portfolios. However, the indirect effect is yet to be observed as financial performance is both affected by direct and indirect effect.

Table VII: multifactor regression of the total effect

VARIABLES RET(=DEP.) ESG 0.00374 (0.0192) DIV 0.868*** (0.196) SIZ -0.317 (0.266) LIQ 13.440 (8.580) LEV -2.960 (2.689) Observations 2,941

Notes: Table VII shows the regression of a multifactor model with dependent variable stock returns as the independent variable. Where ESG stands for ESG, DIV for dividend yield, SIZ for firm size, LIQ for liquidity and LEV for leverage. The longitudinal dataset covers the years 2009-2019.

Implementing a mediating variable allows us to make distinctive conclusions about indirect and direct effects. Therefore, I can focus on the dynamics of direct and indirect effects by implementing operational performance as a mediating variable. The results of the mediation analysis are summarized in TABLE VII. Where the dependent variables model in 3 and 4 are operational-(OPER) and financial performance (RET) respectively. There are no violations regarding heteroskedasticity because of robust corrections to standard errors in SEM. When testing for a normal distribution using the Kurtosis test, a test statistic of 3.373 is obtained, which is significant at a 1 % level meaning that I reject the null-hypothesis of having a

(22)

From the table we see that the conditions for mediation are met since ESG has a significant impact on operational performance(OPER) and operation performance a significant impact on returns(RET). The indirect effect is the product of these coefficients which equals a positive increase of 0.0084(=76.64*0.00011) % in returns. To check whether the indirect effect is different from zero I use the Sobel test (Sobel, 1982). The test statistic of 6.52 means we reject the null-hypothesis of having no indirect effects at a 1 % significance level. Therefore, a 1-point increase in ESG results, via operational performance, in a return increase of 0.0084%. Furthermore, we see that the ESG coefficient shows a direct effect of -0.010%. Comparing this to the effect of ESG found earlier (0.00374%), we observe that the effect of ESG through operational performance neutralizes the direct effect of ESG on returns. This suggests that the finding of better operational performance that flows from an active ESG strategy is already incorporated by the market (Caijas et al., 2011; Eicholtz et al., 2012). As such, I suggest that better ESG performance is observable by the market and excludes arbitrage opportunities stemming from a higher operational performance as a result of an active ESG strategy.

Table VIII: Structural Equation Model results

MODEL 3 MODEL 4

VARIABLES OPER(=DEP.) RET(=DEP.)

OPER 76.64*** (11.74) ESG 0.000111*** -0.010 (3.13e_-03) _(0.020) DIV 0.0352 -1.135*** (0.0318) (0.202) SIZ -0.338*** -0.060 (0.0434) (0.279) LIQ 5.950*** 7.157 (1.400) (8.956) LEV -1.740*** -1.366 (0.440) (2.812) Observations 2,941 2,941

Notes: this table shows the structural equation modelling results that modelled operating performance (OPER) as the mediator variable and the yearly % change in the return index (RET) as the dependent variable. The main independent is ESGwhich is the one-year lagged ESG score, DIV

which is the dividend yield, SIZ which covers the size of the firm computed as the logarithm of the total assets,

LIQ which covers the liquidity and is computed as the total cash divided by the total assets. At last LEV, which covers

the Leverage and is computed as the total cash divided by the total assets. The time span of the mediation analysis

(23)

Conclusion

This paper investigates whether ESG is significantly associated with higher financial

performance using a large number of global listed real estate companies from the period 2009 untill 2019. More specifically, I advance our understanding of the link between ESG and financial performance by shedding light on both the direct and indirect effects. I propose two methods to measure the direct and indirect effects.

First, I analyse the direct effect of ESG and financial performance by constructing portfolios of best- and worst-in-class ESG scores. Accordingly, measuring the returns by the Fama & French 5 factor model I adjust the returns for risk. I find that the market incorporates the integration of ESG in the stock price, since there is no presence of positive significant alphas. Importantly, I also document lower market betas for the best-in-class ESG portfolios. This finding is in line with (Eichholtz, Kok & Yonder, 2012) who suggest that real estate firms with active ESG performance have lower exposure to energy price fluctuations and vacancy risk, and therefore less exposure to the business cycle. Furthermore, the results show that high ESG rated firms are mature firms that, in addition to that, engage in higher excessive

investments compared to their conservative counterpart. This finding indicates that higher investments are necessary when integrating ESG policies.

Second, I propose operational performance as a mediator in the relationship between ESG and financial performance to analyse the indirect effect. I find evidence that active ESG

performance increases higher operational performance and, subsequently, increases the

financial performance. That is, ESG positively affects financial performance through its effect on operational performance. However, since there is no total effect present I conclude that the market already incorporates the value of the effect of ESG on operational performance which is in line with the efficient market hypothesis.

(24)

Reference list

Barnett, M.L. and Salomon, R.M., 2012. Does it pay to be really good? Addressing the shape of the relationship between social and financial performance. Strategic Management

Journal, 33(11), pp.1304-1320.

Blank, H., Sgambati, G. and Truelson, Z., 2016. Best Practices in ESG Investing. The Journal

of Investing, 25(2), pp.103-112.

Cajias, M., & Bienert, S. (2011). Does sustainability pay off for European listed real estate companies? The dynamics between risk and provision of responsible information. Journal of

Sustainable Real Estate, 3(1), 211-231.

Cajias, M., Fuerst, F., McAllister, P. and Nanda, A., 2014. Do responsible real estate companies outperform their peers?. International Journal of Strategic Property

Management, 18(1), pp.11-27.

Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of

finance, 52(1), 57-82.

Chand, M. and Fraser, S., 2006. The relationship between corporate social performance and corporate financial performance: Industry type as a boundary condition. The Business

Review, 5(1), pp.240-245.

Derwall, J., Guenster, N., Bauer, R. and Koedijk, K., 2005. The eco-efficiency premium puzzle. Financial Analysts Journal, 61(2), pp.51-63.

Eichholtz, P., Kok, N. and Yonder, E., 2012. Portfolio greenness and the financial

performance of REITs. Journal of International Money and Finance, 31(7), pp.1911-1929. Eichholtz, P., Kok, N. and Quigley, J.M., 2010. Doing well by doing good? Green office buildings. American Economic Review, 100(5), pp.2492-2509

Fama, E.F. and French, K.R., 2015. A five-factor asset pricing model. Journal of financial

economics, 116(1), pp.1-22.

Hansen, S.D., Dunford, B.B., Boss, A.D., Boss, R.W. and Angermeier, I., 2011. Corporate social responsibility and the benefits of employee trust: A cross-disciplinary

perspective. Journal of business ethics, 102(1), pp.29-45.

Falkenbach, H., Lindholm, A.L. and Schleich, H., 2010. Review articles: environmental sustainability: drivers for the real estate investor. Journal of Real Estate Literature, 18(2), pp.201-223.

(25)

Friede, G., Busch, T., & Bassen, A. (2015). ESG and financial performance: aggregated evidence from more than 2000 empirical studies. Journal of Sustainable Finance &

Investment, 5(4), 210-233.

Giacomino, D.E. and Mielke, D.E., 1993. Cash flows: Another approach to ratio analysis. Journal of Accountancy, 175(3), p.55.

Giese, G., Lee, L.E., Melas, D., Nagy, Z. and Nishikawa, L., 2019. Foundations of ESG Investing: How ESG Affects Equity Valuation, Risk, and Performance. The Journal of

Portfolio Management, 45(5), pp.69-83.

Ioana, D. and Adriana, T., 2013. New corporate reporting trends. Analysis on the evolution of integrated reporting. Annals of the University of Oradea, Economic Science Series, 22(1), pp.1221-1228.

John, T.A., 1993. Accounting measures of corporate liquidity, leverage, and costs of financial distress. Financial Management, pp.91-100.

Kempf, A. and Osthoff, P., 2007. The effect of socially responsible investing on portfolio performance. European Financial Management, 13(5), pp.908-922.

Kivits, R. A., & Furneaux, C. (2013). BIM: enabling sustainability and asset management through knowledge management. The Scientific World Journal, 2013.

Lee, D.D., Faff, R.W. and Rekker, S.A., 2013. Do high and low-ranked sustainability stocks perform differently?. International Journal of Accounting & Information Management, 21(2), pp.116-132.

Little, T.D., Card, N.A., Bovaird, J.A., Preacher, K.J. and Crandall, C.S., 2007. Structural equation modeling of mediation and moderation with contextual factors. Modeling contextual

effects in longitudinal studies, 1, pp.207-230.

Lindholm, A.L. and Leväinen, K.I., 2006. A framework for identifying and measuring value added by corporate real estate. Journal of Corporate Real Estate, 8(1), pp.38-46.

Liow, K.H., 2010. Firm value, growth, profitability and capital structure of listed real estate companies: an international perspective. Journal of Property Research, 27(2), pp.119-146. Lizieri, C., 2013. After the fall: Real estate in the mixed-asset portfolio in the aftermath of the global financial crisis. The Journal of Portfolio Management, 39(5), pp.43-59.

Lougee, B. and Wallace, J., 2008. The corporate social responsibility (CSR) trend. Journal of

Applied Corporate Finance, 20(1), pp.96-108.

Luo, X., & Bhattacharya, C. B. (2009). The debate over doing good: Corporate social performance, strategic marketing levers, and firm-idiosyncratic risk. Journal of

Marketing, 73(6), 198-213.

(26)

Nafukho, F.M., Hairston, N. and Brooks, K., 2004. Human capital theory: Implications for human resource development. Human Resource Development International, 7(4), pp.545-551. Petersen, H.L. and Vredenburg, H., 2009. Morals or economics? Institutional investor

preferences for corporate social responsibility. Journal of Business Ethics, 90(1), p.1. Pivo, G., 2008. Responsible property investment criteria developed using the Delphi Method. Building Research & Information, 36(1), pp.20-36.

Ribando, J.M. and Bonne, G., 2010. A new quality factor: Finding alpha with ASSET4 ESG data. Starmine Research Note, Thomson Reuters, 31.

Ruf, B. M., Muralidhar, K., Brown, R. M., Janney, J. J., & Paul, K. (2001). An empirical investigation of the relationship between change in corporate social performance and financial performance: A stakeholder theory perspective. Journal of business ethics, 32(2), 143-156.

Sassen, R., Hinze, A.K. and Hardeck, I., 2016. Impact of ESG factors on firm risk in Europe.

Journal of Business Economics, 86(8), pp.867-904.

Sobel, M.E., 1982. Asymptotic confidence intervals for indirect effects in structural equation models. Sociological methodology, 13, pp.290-312.

Supanti, Daraneekorn, Ken Butcher, and Liz Fredline. "Enhancing the employer-employee relationship through corporate social responsibility (CSR) engagement." International

Journal of Contemporary Hospitality Management 27, no. 7 (2015): 1479-1498.

Surroca, J., Tribó, J.A. and Waddock, S., 2010. Corporate responsibility and financial performance: The role of intangible resources. Strategic management journal, 31(5), pp.463-490.

Tamimi, N. and Sebastianelli, R., 2017. Transparency among S&P 500 companies: An analysis of ESG disclosure scores. Management Decision, 55(8), pp.1660-1680.

Waddock, S., 2008. Building a new institutional infrastructure for corporate responsibility. Academy of Management perspectives, 22(3), pp.87-108

Wagner, M., & Schaltegger, S. (2003). How does sustainability performance relate to business competitiveness?. Greener Management International, 5-16.

Wimmer, M., 2013. ESG-persistence in socially responsible mutual funds. J. Mgmt. &

Sustainability, 3, p.9.

(27)

Appendix A: Correlation matrix of variables in the mediation analysis

Notes: This table summarizes the time average of panel data correlations ranging from the period 2009-2019. The tables concedes of the variables yearly change in return index (RET), operational performance (OPER), the ESG score (ESG), the liquidity (LIQ), the leverage (LEV) and the dividend yield (DIV).

Appendix B: Sobel Test (1982)

]*_ _`_*ab_c`_C]`_*ab

d`

=z (4)

Appendix C: Performance level portfolios in a multifactor regression model, 2009-2019 (ENGX as market premium)

α 0.00276 0.00838** 0.00206 0.00540* 0.00437 0.00632 0.00103 (0.00508) (0.00418) (0.00593) (0.00316) (0.00660) (0.00517) (0.00653) ß0 (ß) 0.670*** 0.444*** 0.674*** 0.634*** 0.632*** -0.230** 0.00214 (0.110) (0.110) (0.125) (0.0908) (0.142) (0.107) (0.157) ß1 (SMB) -0.724** -0.463 -0.646 -0.244 -0.870* 0.184 0.627 (0.355) (0.303) (0.428) (0.271) (0.520) (0.367) (0.532) ß2 (HML) -0.915** 0.0423 -1.222** -0.193 -1.491** 1.264** 1.298** (0.447) (0.422) (0.527) (0.359) (0.575) (0.507) (0.574) ß3 (RMW) -1.295** -1.487*** -1.210* -1.045** -1.802** -0.277 0.758 (0.573) (0.502) (0.682) (0.460) (0.832) (0.620) (0.800) ß4 (CMA) 0.772 -0.241 1.103** -0.458 1.133* -1.344* -1.591** (0.476) (0.535) (0.552) (0.419) (0.678) (0.679) (0.714) Observations 120 120 120 120 120 120 120 R-squared 0.306 0.301 0.239 0.422 0.192 0.135 0.059

and ESG-25 consist of the highest and lowest rated 25 % ESG rated real estate firms in the dataset. The ESG10 and ESG-10 are

the highest and lowest rated ESG rated real estate firms based on a 10 % level. The ESG∆25 is the differenced portfolio the 25

% ESG robust standard errors are between brackets reported under the applicable coefficients. The time span is based on 2009-2019. ***, **, * indicate significance levels based on 1, 5 and 10 % level.