ESG in investor profiling: what are we talking about? (2024)

2.1 Gender

Studies on the relationship between investor gender and attitude to ESG investment are mainly convergent in their findings. Although Lewis and Mackenzie (2000) found that the major investors in a UK ethical unit trust were male, Roberts (1996) and Laroche etal. (2001) showed that women are more socially responsible than men. Tippet and Leung (2001) confirmed that women seem readier to invest in socially responsible financial products. In the same vein, Junkus and Berry (2010) argued that the typical socially responsible investor is female, a result confirmed by Diouf etal. (2016). Beal and Goyen (1998) and Junkus and Berry (2010) found that women tend to invest a greater portion of their mutual fund investment portfolio in SRI-profiled mutual funds than men. Based on this previous research, we formulate the following hypothesis:

2.2 Age

Rosen etal. (1991) and Tippet and Leung (2001) argued that socially responsible investors are mainly young people. Analogously, Hayes (2001) found that people aged between 18 and 34 have a stronger inclination towards ESG investments. Starting from an interdisciplinary review of the literature, Diamantopoulos etal. (2003) tested a large nationwide sample of British consumers. They demonstrated that young people are more socially oriented than older people. Similarly, using data from a survey of the retail clients of two Netherlands banks specialized in the provision of exclusively socially responsible investments, Bauer and Smeets (2015) demonstrated that young people identify more with socially responsible investments compared to others. McLachlan and Gardner (2004) introduced a distinction between conventional and ethical investors, supporting the idea that younger investors are more socially responsible than others. Woodward (2000), Nilsson (2008), and Junkus and Berry (2010) argued that younger investors invest a greater portion of their mutual fund investment portfolio in SRI-profiled mutual funds than older investors. Conversely, D’hondt etal. (2022), using both survey and trading data from 9,286 retail investors between 2005 and 2011, maintained that the older the investor, the higher their exposure to the three ESG factors. The same conclusion was drawn by Hafenstein and Bassen (2016), and again by Rossi etal. (2019), who used a specific questionnaire administered to a representative panel of Dutch households. Lewis and Mackenzie (2000), using a specific survey administered to the investors in a UK ethical unit trust, found that the majority are over 45years of age. Bearing in mind these mixed results, we formulate the following hypothesis:

2.3 Education level

The studies by McLachlan and Gardner (2004) and D’hondt etal. (2022) showed that highly educated investors present low ESG scores in the allocation of their portfolio assets. In particular, Anderson and Robinson (2022) concluded that investors with strong pro-environmental behaviors and beliefs are generally uninterested in financial matters. However, the majority of the literature on this topic asserts the opposite. Rosen etal. (1991), Woodward (2000), Tippet and Leung (2001), Haigh (2008), Nilsson (2008), Junkus and Berry (2010), and Bauer and Smeets (2015), among others, argued that better-educated investors invest a greater portion of their mutual fund investment portfolio in SRI-profiled mutual funds than less educated investors. Riedl and Smeets (2017) and Rossi etal. (2019) found that individuals with university degrees are more likely to invest in SRI funds than less educated investors. Based on this literature review, we formulate the following hypothesis:

2.4 Financial literacy and financial knowledge

Previous works (e.g. Chen and Volpe, 1998; Lusardi and Mitchell, 2007, 2008, 2014) have considered financial literacy and financial knowledge to be interchangeable and we follow this approach in this paper. Few studies exist on the relationship between financial literacy and ESG portfolio choices. Gutsche etal. (2021) found a positive relationship between financial literacy and attitude toward sustainable investments. Rossi etal. (2019), Anderson and Robinson (2022), and D’hondt etal. (2022) found a negative relationship between financial literacy and ESG portfolio scores. Varmaz etal. (2022) showed that financial literacy is negatively associated with students’ choice of ESG investment funds, although financial literacy is not significant. Based on this contradictory literature review, we formulate the following hypothesis:

2.5 Employment status

Very few studies have considered employment status as a variable that influences investment decision-making with respect to ESG financial products. Rosen etal. (1991) compared two groups of investors in mutual funds: socially responsible investors versus “standard” investors. They considered age, marital status, income, educational level, and employment status. They found that 81% of socially responsible investors were white-collar workers, 6.5% were blue-collar workers, and 12.5% were retired. On the other hand, 63% of “standard” investors were white-collar workers, 17% were blue-collar workers and 20% were retired. Despite this segmentation, the authors did not provide statistical evidence of the role played by the investors’ employment status in their decision to be a sustainable or a “standard” investor. Based on the statistics presented above, they simply argued that socially responsible investors are more likely to hold white-collar jobs.

Again, Lewis (2001) compared seven groups composed of “standard” investors with seven groups of “ethical or green” investors. Using a qualitative methodology, the author argued that a majority of “ethical or green” investor groups worked in caring professions such as health, teaching, and nursing, thus confirming the conclusion reached by Rosen etal. (1991). However, a bias is identifiable in the composition of these groups, because not all kinds of employment status were considered. For example, there were no retired people in the “ethical or green” groups while only one retired soldier was included in the “standard” groups. Based on these very few studies, we formulate the following hypothesis:

3.1 The questionnaire and data collected

Using a questionnaire to profile socially responsible investors is common practice. Through an online survey, Pérez-Gladish etal. (2012) considered a sample of 145 investors in Australia to profile socially responsible investors. Valor etal. (2009) considered a sample of 400 telephone interviews with Spanish citizens conducted by a specialized market research agency to quantify the demand for socially responsible investments. McLachlan and Gardner (2004) compared a sample of 55 conventional investors and a sample of 54 ethical investors in Australia using a postal questionnaire. In the same vein, we also use a questionnaire. However, the difference between our study and the others is that the questionnaire used in this analysis was not created by us but is officially recognized by the relevant Italian Securities and Markets Supervisory Authority (CONSOB) as meeting the MiFID requirements. It contains, therefore, socio-demographic questions (gender, age, educational level, job position), questions aimed at determining the investor’s financial literacy knowledge, including their investment time horizon, and questions relating to their interest (or lack thereof) in ESG issues, to facilitate building a portfolio assetallocation that adequately considers these aspects. We highlight that the data used in this analysis comes from the questionnaire legally recognized by the relevant authority and that the responses collected are used by banks for customer risk profiling as well as for assessing the consistency of the subscribed investments with the view of sustainability professed.

To provide more details in relation to the responding investors’ preferences in terms of ESG objectives, the investors could choose between ESG goals that are relevant to their specific interests, indicating one or more objectives that could characterize their investments. Based on the data provided to us, we observe that some respondents expressed their interest and attitude as relating solely to the environmental pillar, others (although very few) solely to the social pillar and no one expressed interest purely in the governance dimension. Some expressed interest in the combination of the environmental and social pillars (no other pairings were observed), while others expressed interest in all threeE, S and G pillars jointly. These observations allowed us to construct our dependent variable (ESG_attitude) as an ordinal variable that can be assigned a value from 1 to 3, where 1 relates to those customers who are exclusively interested in solely one pillar (environment or social), 2 identifies those investors who are interested in two of the three ESG pillars (specifically environment and social issues), while 3 identifies those individuals who are interested in investing their savings in financial products that pursue all three ESG components without distinguishing between them. Among the independent variables employed in this empirical study, we consider the gender of investors and their age. The latter is a continuous variable (age) with its square (age2) introduced to account for possible non-linearities that can be interpreted from the contradictory results presented in the literature review (Hallahan etal., 2004; Grable etal., 2006; Faff etal., 2009). To capture the educational attainment of respondents we employed three dummy variables, assuming a value equal to 1 to correspond to the highest level of education held by individuals and 0 otherwise (Table1). The variable financial literacy (FL) was measured in additive terms, using the sum of the values assigned to respondents’ answers to three questions included in the questionnaire which proxy the degree of investors’ financial literacy. The first of the three questions concerned the frequency with which respondents obtain information on financial markets’ performance. In answering this question, customers could indicate that they collect information regularly, in which case we assigned a value of 2; that they do this monthly, in which case we assigned a value of 1; or, finally, that they do not obtain information at all, for which we assigned a value of 0. The second question asked the client to indicate whether the riskiest investments are generally more profitable, requiring a yes or no response. By answering affirmatively, customers demonstrate that they are aware of the relationship between risk and return on financial investments, with the opposite being the case for negative answers. We assigned a value of 1 for a positive answer and 0 for a negative. The third question required the client to assign a level of risk, either low or high, to three financial instruments (specifically, derivatives, stocks, and money market funds). As is well-known, stocks and derivatives present a high risk, and money market funds a low risk. Each of the correct answers was assigned a value equal to 1, with 0 being assigned in the case of incorrect answers. Therefore, the additive variable constructed as a proxy of financial literacy could assume values ranging from 0 to 6, with the maximum value coinciding with the maximum level of financial literacy (see Table1).

Respondents’ employment status is also proxied with dummy variables (see Table1). Finally, we introduced two original control variables: investment horizon (LT) and preference for splitting wealth between banks (diversification) (see Table1). The investment horizon, and in particular the investor’s long-term focus (Pérez-Galdish, 2012) is crucial in determining the assetallocation strategy for their portfolio, particularly in ESG investments. In the literature, a relationship is identified between SRI returns and a long time horizon (e.g. Sørensen and Pfeifer, 2011; Fuller, 2012; Wagemans etal., 2013). In line with the prevailing literature (e.g. Dorfleitner and Utz, 2014) we considered short-term as less than one year, medium-term as from one to five years, and a long-term horizon as being more than five years. Thus, respondents with short-term portfolios were assigned an LT value of 0, those with medium-term an LT value of 1, while those with long-term orientation obtained an LT value of 2. The second control variable measures the tendency towards diversification between banks. The variable diversification is a dummy that takes the value of 1 if the respondents use more than one bank to manage their financial wealth and 0 if all their wealth is managed by a single bank.

For each variable used in this study, we provide a description and our expectation as to its ESG attitude (Table1).

3.2 Descriptive statistics

Each investor in our dataset declared themselves interested in investing in ESG financial products but in different ways. Some investors (21.04% of the sample) considered only one pillar (more specifically, 16.32% considered only the environmental aims and 4.74% considered only the social dimension), others (12.11%) took into consideration a combination of environmental and social goals, while 71.58% of investors considered all three ESG aims in their decision-making on assetallocation.

Table2 shows some descriptive statistics of the overall sample (N=190) and highlights the specific features of the panel of available data. Tables 3, 4, 5, and 6 present some descriptive statistics related to respectively environmental investor profile (n=31; 16.32% of the sample), social investor profile (n=9; 4.74% of the sample), environmental and social investor profile (n=14; 7.37% of the sample), and environmental, social, and governance investor profile (n=136; 71.58% of the sample).

As shown in Table2, the average age of the investors was 51.54years. The sample is fairly balanced between men (53.68%) and women (46.32%) and consists of 44.74% employees, 11.58% self-employed workers, 12.11% entrepreneurs, 18.95% retired workers, 5.26% students, and 7.37% unemployed. People with a university degree or equivalent represented 23.68% of the sample, those who only completed high school represented 53.16%, and 23.16% of investors reported a low level of education. In our sample, 57.37% of investors concentrated their entire financial wealth in a single bank and 42.63% spread their savings between several banks.

Finally, the level of the respondents’ financial literacy was quite high; the possible scores using our methodology range from 0 to 6 but in our sample, the range is between 4 and 6.

Table7 shows the matrix correlation.

4.1 Step 1 – profiling investors according to their single or multiple ESG attitude

Since ESG_attitude is an ordinal variable ranging from 1 to 3, an ordered logistic regression was used. The estimated equation is as follows:

• •

  gender indicates if the investor is female (0) or male (1);

• •

  age and age2 indicate respectively the age and its squared value for each investor;

• •

  edu_i indicates the education level held by the investor, with specifically edu_1 representing a low level of education, edu_2 a medium level of education, and edu_3 a high level of education;

• •

  FL proxies the financial literacy score obtained by each investor;

• •

  job_k indicates whether the individual was unemployed (nojob), a student, self-employed, employed, an entrepreneur, or retired;

• •

  LT indicates the investor’s short-, medium- or long-term orientation;

• •

  diversification indicates whether the investor has concentrated all their financial wealth in a single bank or spread their savings between several banks.

The estimates are reported in Table8.

Our results suggest that female investors pay more attention to the three ESG aims collectively while male investors seem to have a more pronounced interest in one pillar only. Our H1 is therefore confirmed. Age significantly influences the investor’s ESG attitude. However, the significance and sign of the variable “age-squared” show that the relationship between age and ESG attitude is not necessarily linear. More specifically, we can argue a positive relationship between age and the three ESG aims in those over 44years old while younger investors (under 44years old) tend to pay attention to just one element of sustainability. Based on this result, our H2 is confirmed. Investors with a high level of education expressed more interest in environmental, social, and governance issues than those with a school-leaving diploma (a medium level of education). Thus our H3 is confirmed. However, financial literacy is not found to be statistically significant; thus our H4 is rejected.

Retired investors are more interested in the three ESG targets than people without a job, thus H5 is rejected. This result cannot be explained by the level of education of the retired investors in our sample. Indeed, only 2 retirees out of the 36 in the sample (6%) held a university degree. Therefore, this focus on all ESG goals could be an expression of other aspects not controlled for, such as, for example, a greater concern for leaving a better world to their descendants.

The long-term investor orientation is not statistically significant in our analysis.

One very interesting result however is that investors who concentrate their portfolio in a single bank are more oriented towards the three sustainable goals (environment, social, and governance) collectively than those who diversify their assets between several banks.

4.2 Step 2 – disentangling ESG investor’s profile

Several studies (e.g. Beal and Goyen, 1998; Woodward, 2000; Nilsson, 2008; Junkus and Berry, 2010) have attempted to define the profile of the socially responsible investor. In more recent years, Bethlendi etal. (2022) examined heterogeneity among consumers of green finance and the possibility of identifying different investors' sustainable profiles according to their socio-demographic characteristics. Intending to verify this statement, whereas in step 1 we performed a general analysis of the answers provided by the sample examined, in step 2 we aimed to draw a more defined socio-demographic profile of the investor who is more oriented towards one or other of the E and S pillars (as no one expressed interest solely in the governance pillar), or to a combination of E and S, or to E, S, and G jointly. To this purpose, we introduce a dummy variable (d_ESG_attitude) which takes a value of 1 or 0 according to the following scheme. If the respondent expressed interest solely in either the environmental or the social pillar, the d_ESG_attitude takes a value of 1; otherwise 0. If the investor expressed interest in a combination of E and S jointly, the d_ESG_attitude takes a value of 1; otherwise 0. Finally, if the respondent expressed interest in all threeE, S and G pillars jointly, the d_ESG_attitude takes a value of 1; otherwise 0. Thus, we repeat (1) four times using a logit regression. In this way, we obtain four theoretical ESG investor profiles as shown in Table9.

The results presented in Table9 appear to be in line with those shown in Tables 8 and in Appendix. Moreover, they allow for the identification of a specific ESG investor profile. First of all, the results shown in Table9 column (b) are weak (Prob>χ² 0.3416) and without statistical significance, therefore it is not possible to profile the investor who is oriented solely towards the social pillar. This is doubtless due to there being only a small number of investors in our sample who expressed their interest as being in favor of the social theme alone. The results shown in columns (a), (c), and (d) of Table9 are statistically significant (Prob>Chi2 0.0004; Prob>Chi2 0.0153; Prob>Chi2 0.0003 respectively). However, column (c) reports only two significant variables: age and long-term orientation. Therefore we can argue that investors who expressed an interest in environmental and social issues jointly are young people (under the age of 46) and do not have a long-term orientation. This last result represents a contradiction since long-term orientation is an essential element for achieving sustainability aims. Columns (a) and (d) demonstrate the greatest significance.

Comparing columns (a) and (d) of Table9, one of the main results is that men tend to be more environmentally oriented than women who, by contrast, are more oriented towards the environment-social-governance issues jointly. Again, our H1 is confirmed.

Investors who are over 44years of age are more interested in the ESG pillars jointly (again confirming our H2) while investors under 44 seem to be more interested in environmental goals alone. The level of education is significant only in column (d), highlighting that investors with a medium level of education are less interested in the threeE S and G pillars jointly than those with a high level of education. Thus, again, our H3 is confirmed. Retirees appear to be more interested in all three ESG pillars considered jointly than those who are entrepreneurs. By contrast, investors who are unemployed seem to be less interested than entrepreneurs in the ESG pillars jointly. Thus, again, our H5 is rejected. In Table9, financial literacy is not statistically significant in any column, thus, again, our H4 is rejected. Investors who express a greater interest solely in environmental aims are more likely to have relationships with multiple banks. Conversely, those who report an interest in all three ESG pillars jointly tend to have a higher concentration of portfolios with a single bank. Again, the long-term orientation is not statistically significant.