Intervention (Hidden)
Recent studies show that different rating agencies often rate the sustainability level of the same company very differently (Berg et al. 2022). This could unsettle investors looking for sustainable investments, as the inconsistency of sustainability ratings can create uncertainty about the sustainability of financial products and thus reduce the attractiveness of sustainable investments (Avramov et al. 2021). Our project investigates how private investors make their investment decisions depending on available sustainability ratings and whether they react to the inconsistency in sustainability ratings in their investments.
Experimental design: We plan to conduct a survey among German private investors, aiming for a representative sample of this demographic. The survey will include an investment experiment in which respondents allocate a hypothetical sum of money (€1,000) between two funds: a conventional fund and a sustainable fund. For each fund, we will display two sustainability ratings, sourced from two different rating agencies. These ratings will be presented as sustainability percentile ranks, assigned by the respective agencies to the sustainable fund. All displayed characteristics of the funds across all choice sets, aside from the sustainability ratings of the sustainable fund, will be identical. In each allocation task, participants will be shown a unique combination of sustainability ratings for the sustainable fund. There will be nine different combinations of sustainability ratings for the sustainable fund that participants will use to make their allocation decisions: (25, 25), (25, 50), (25, 75), (50, 25), (50, 50), (50, 75), (75, 25), (75, 50), and (75, 75). These two numbers in parentheses represent the percentile ranks compared to other funds in terms of sustainability. For example, (75, 50) indicates that one agency places the fund in the 75th percentile, while another places it in the 50th percentile of sustainability performance. The variation reflects how diverging ratings may influence investment decisions, with the hypothesis being that larger divergences in ratings reduce investments in the sustainable fund. The ratings of the conventional fund will remain constant at (50, 50) across all choice sets.
Each respondent is asked to make all allocation decisions. All participants begin with the decision where both ratings for the sustainable fund are 50. The order in which participants are presented with the other eight combinations of sustainability ratings is randomized. Additionally, the order in which the rating agencies' names (MSCI ESG and Refinitiv) appear in the table of fund information is randomized across participants. Respondents are asked to indicate how much money they would allocate to the sustainable and conventional funds. Whether the conventional or the sustainable fund is listed first is also randomized across participants.
The dependent variable is the proportion of money invested in the respective sustainable fund for each combination of sustainability ratings. In order to test investment preferences based on sustainability ratings, we will compare the investment decisions that participants make when exposed to different rating combinations. In the baseline regression, we will regress the share of investments in the sustainable fund on the mean of both ratings and the absolute difference between the ratings. Since retail investors are expected to have some preferences for sustainable investments, we anticipate that the coefficient for the mean rating will be positive. If investors are sensitive to rating divergence, we expect the coefficient for the absolute difference between the ratings to be significant and negative, indicating that investors dislike uncertainty about the sustainability of their investments. We may conduct additional regressions to address nonlinearities that arise when the relationship between investments and the mean rating is concave.
In another set of regressions, we use dummy variables to test for divergence aversion in sustainability ratings. We do this by regressing the investments in the sustainable fund on dummies indicating whether each of the sustainable funds’ two ratings is either below or above 50, resulting in four dummy variables. Additionally, we include a dummy indicating whether there is a difference between the two ratings. We expect this dummy to have a significant and negative coefficient, suggesting that retail investors penalize discrepancies in ratings with lower investments in the sustainable fund. If our results show sufficient statistical power, we will also differentiate between small rating differences of 25 and large differences of 50.
Our survey includes questions on whether participants have heard of ESG ratings and have preferences toward certain rating agencies. We examine the awareness and preferences of retail investors regarding ESG rating agencies, thereby contributing to a more nuanced understanding of how ESG ratings are perceived among investors. We may include additional regressions to test whether individuals respond more strongly to ratings from agencies they trust. The inclusion of these regressions will depend on the statistical power of our results and the number of respondents who express preferences for specific rating agencies.
The survey will also include questions on the ownership of sustainable assets, risk aversion, sustainability preferences, expectations and opinions on sustainable assets, and the aspects of investments that investors generally focus on. Additionally, we will ask questions related to financial literacy and sustainable finance literacy (Filippini et al. 2024).
These observables will enable us to further investigate what motivates individuals to change their allocations when ratings differ. Therefore, we plan to run heterogeneity analyses to determine which groups of retail investors react more to higher or lower ratings. Individuals with stronger sustainability preferences, higher return expectations of sustainable assets, have a better understanding of sustainable finance or report greater trust in the accuracy of ratings may increase their investments in the sustainable fund more strongly if either of the ratings is increased. Also, some groups should react more strongly to rating divergence. For example, investors with greater risk aversion or those who prefer certainty about the sustainability of an asset should reduce their sustainable investments more in the face of divergent ratings than others.
Sample: We plan to conduct a survey among 2,000 German retail investors. To achieve a sample that is representative of this population, we aim for the following age distribution: approximately 30% of participants aged 18-39, 20% aged 40-49, 20% aged 50-59, and 30% aged 60 and older. Additionally, we aim for the sample to consist of 40% females.
Main ex-ante hypotheses: We expect retail investors to allocate more funds to a sustainable investment when the sustainability ratings are higher. Conversely, lower ratings will result in reduced investment. When sustainability ratings differ significantly between agencies, we anticipate that investors will reduce their investment in the sustainable fund due to increased uncertainty regarding its sustainability.
References:
Avramov, D., Cheng, S., Lioui, A., & Tarelli, A. (2022). Sustainable investing with ESG rating uncertainty. Journal of financial economics, 145(2), 642-664.
Berg, F., Koelbel, J. F., & Rigobon, R. (2022). Aggregate confusion: The divergence of ESG ratings. Review of Finance, 26(6), 1315-1344.
Filippini, M., Leippold, M., & Wekhof, T. (2024). Sustainable finance literacy and the determinants of sustainable investing. Journal of Banking & Finance, 163, 107167.