Professional Documents
Culture Documents
Esg Rating and Stock
Esg Rating and Stock
Abstract
For example, in the older company ratings, Royal Dutch Shell and Microsoft
both scored 75 out of 100 within their industry groups, ranking in the best
quartile among their peers. (Higher scores were better in this version.) These
were good scores, indicating the two companies were among the best-in-class
ESG performers in their respective peer groups. At the portfolio level, assum-
ing their position size was the same, they would have had the same impact
on the Portfolio Sustainability Score.
In the new company ESG Risk rating, by contrast, Royal Dutch Shell’s score is
34, an indicator of High ESG Risk, while Microsoft’s score is 13.8, an indicator
of Low ESG Risk. While both companies do reasonably well managing the
material ESG risks they face relative to their peers–one reason why their old
scores were similar–Royal Dutch Shell operates in an industry that carries
far more ESG risk exposure. As a result, its ESG Risk rating has a much
more negative impact on the Portfolio Sustainability Score compared with
Microsoft’s rating.
That said, the new rating remains sensitive to best-in-class comparisons. For
example, a portfolio that holds an oil company would be better o↵ with expo-
sure to Royal Dutch Shell, with its ESG Risk rating of 34, than with exposure
to ExxonMobil, which has an ESG Risk rating of 40.5.
ing – https://www.morningstar.com/articles/954595/enhancement-to-sustainability-rating-emphasizes-
material-esg-risk.
16
This discussion only briefly describes a more complicated methodology as we do
not focus on the specific information content of the new ratings. For more in-
formation on how the ESG Risk ratings are calculated, see https://globalaccess-
tutorials.s3.amazonaws.com/ESG%20Risk%20Rating Methodology%20document.pdf.
17
https://www.morningstar.com/articles/954595/enhancement-to-sustainability-rating-emphasizes-
material-esg-risk.
10
11
We collect ESG ratings from two sources for all companies trading on the NYSE, NAS-
DAQ, and Amex exchanges between June 2019 and January 2020. First, we obtain both
the old ESG rating and the new ESG Risk ratings from Morningstar Direct and confirm
12
13
14
15
16
17
18
19
4 Empirical Results
Our empirical strategy is to test the di↵erences in abnormal returns and ownership sur-
rounding the adoption of the new ESG Risk rating by Morningstar and Yahoo! Finance.
If ESG ratings are salient to investors but some investors are unsophisticated, the change
in methodology may result in misinterpretation about its meaning. In other words, in-
vestors who rely on the ratings available through Morningstar and Yahoo! Finance to
provide them with information about a firm’s ESG exposure but who are unlikely to do
their own due diligence, may mistakenly interpret the inversion of the rating scale in the
wrong way. In this case, we expect that retail investors, who are assumed to be otherwise
uninformed, will interpret a decline in the ESG rating after adoption as bad news and an
increase as good news even when this is not the case.
More sophisticated and informed investors, such as 13F institutions and ESG funds
will either be aware of the implications of the change or will have already previously
incorporated the change through their subscription to Sustainalytics. Finally, short sell-
ers, who are also informed, may take advantage of the uninformed investors confusion
and trade accordingly. Therefore, if enough investors are confused and rebalance their
portfolio based on their confusion, then we expect that abnormal returns will be negative
when the rating declines and positive when the rating increases even after controlling for
any new information that the change in rating may contain. Moreover, we predict that
this e↵ect will be primarily driven by the changes in participation by uninformed or retail
investors and holdings of short sellers, but not by the change in holdings of institutional
investors such as those that report on Form 13F or funds whose primary objective is
sustainable investing.
20
We begin our analysis by examining the e↵ect of the ESG rating methodology change on
the firm’s abnormal returns. We estimate the following di↵erence-in-di↵erences specifi-
cation:
where AReti,t is stock i’s abnormal return in month t computed with either a single-,
three-, or four-factor model. ESGi captures stock i’s exposure to the change in the
firm’s rating due to the adoption by Morningstar and Yahoo! Finance of the new Sustain-
alytics ESG Risk rating. We measure the change in the ESG rating in four ways: as the
change between September 2019 and October 2019 (calculated as the di↵erence between a
firm’s ESG Risk rating in October 2019 and its ESG rating in September 2019), ESGi ,
and as the average change over three month periods after and three months before the
adoption, ESGi . We also use two indicator variables to capture the direction of the
ESG rating change: ESG Pseudo-Downgradei and ESG Pseudo-Upgradei . di
and dt represent firm and month fixed e↵ects. Standard errors are clustered at a firm
level.23
If our conjectures about the e↵ect of investor confusion are correct, then we expect
the coefficient on the interaction terms that include our measures of ESG rating changes
to be positive and significant, i.e., a decline (increase) in ESGi would induce investors
to sell (buy) the stock causing negative (positive) abnormal returns from October 2019
onward. We report the regression estimates of equation (1) in Table 3. Panel A presents
the baseline regression. Single-factor abnormal returns are presented in columns (1) to
(4), three-factor abnormal returns are presented in columns (5) to (8), and in the last
23
Our time-series consists of only six months around the methodology adoption. According to Angrist
and Pischke (2008), six clusters is not enough for the standard cluster adjustment. See, Chapter 8.2.3
‘Fewer than 42 Clusters’ of Mostly Harmless Econometrics for more details.
21
22
In order to disentangle the inversion of the scale from new information about a firm’s ESG
risk exposure conveyed by the change in methodology, we control for two possible types
23
24
25
The findings of the previous section indicate that investor reliance on ESG ratings and
their confusion about the new methodology impacts stock prices. This impact, however,
should be short-lived and ultimately reversed as investors recognize their error and reverse
their positions and/or market forces correct it. To test whether this is the case, we
estimate a dynamic version of the regression equation (1) as follows (Célerier and Matray,
2019):
5
X
AReti,t = 0 + e ESG Pseudo-Downgradei ⇥ d(e)t + di + dt + "i,t , (2)
e= 4, e6= 1
where AReti,t is the abnormal four-factor monthly return and d(e)t is equal to one
exactly e periods after (or before if e is negative) the implementation of the new ESG
rating methodology. We use September 2019 as the reference month. di and dt denote
stock and year-month fixed e↵ects and standard errors are clustered by firm. The main
coefficient of interest is e, that captures the di↵erential e↵ect over time for firms that
experience the most severe decline in the value of their ESG rating (treated firms where
ESG Pseudo-Downgrade=1), against those that have less severe declines (control
firms where ESG Pseudo-Downgrade=0). We expect that the di↵erence in abnormal
26
27
The misinterpretation of some investors about the new methodology results in price pres-
sure as they sell firms that experience a decline in their ESG rating and buy firms that
experience an increase. Since confusion is more likely among uninformed investors, we an-
ticipate that changes in ESG ratings will be associated with a change in the participation
of retail investors. Sophisticated investors, such as institutional investors and ESG funds,
are less likely to misunderstand the implications of the ESG rating change or have already
incorporated the change when Sustainalytics revised the methodology in September 2018.
Therefore, we expect that these investors will not rebalance their portfolio in response
to the adoption of the new ratings regime by Morningstar and Yahoo! Finance. Finally,
since Morningstar announced that it would adopt the new ESG Risk rating in July 2019,
informed investors such as short sellers, may take advantage of the confusion of less in-
formed investors. (It is also possible, of course, that institutional investors may similarly
take the other side of uninformed investors’ trades.) In our analysis, we consider whether
retail participation, institutional and ESG fund ownership, and short interest change in
relation to the change in the firm’s ESG rating.
We follow Mian and Sufi (2011) and estimate the generalized di↵erence-in-di↵erences
specification of the form:25
where Ownershipi is defined as the change from the quarter after to the quarter before
the adoption in 1) the average number of investors in a firm on the Robinhood platform,
2) the average percentage of shares outstanding held by 13F investors, 3) the average
25
This specification is equivalent to Mian and Sufi’s (2011) first-stage regression from equation (2).
We replace HouseP riceGrowth0206zm with Ownershipi on the left hand side of the equation and
Elasticitym,1997 with ESGi on the right hand side of the equation.
28
29
30
31
The results of the prior sections suggests that confused investors, who value firms’ sus-
tainability, may impact prices and these investors are more likely to be uninformed retail
investors. In this section, we further examine whether the e↵ect of confused investors on
abnormal returns is attenuated in firms that have large institutional holdings compared
to all other firms. Retail traders in firms with high institutional ownership are unlikely
to be the marginal investors and to move prices to the same degree as firms with greater
retail participation. In addition, information is more widely available to investors in these
firms. We, therefore, expect that firms with higher institutional holdings will not be as
a↵ected by the adoption of the new methodology by Morningstar and Yahoo! Finance as
firms with lower institutional holdings.
In Table 7, we examine abnormal returns using a triple di↵erence-in-di↵erences regres-
sion where our main independent variable of interest is ESGi ⇥ Postt ⇥ High Insti .
High Insti is an indicator variable equal to one if the stock’s ownership is in both the
highest quartile of institutional ownership and the lowest quartile of retail participation,
and zero otherwise. We also control for any change in the Sustainalytics ESG Risk rating
over time as well as any new information conveyed by the change in methodology through
the change in the firm’s relative ranking.
The coefficient on the triple interaction term is negative and significant indicating
that the e↵ect of investor confusion on abnormal returns is lower for firms that have
high institutional ownership but low retail participation compared to firms that have
lower institutional ownership and higher retail participation. As a robustness test in
untabulated results, we perform the same analysis but substitute an indicator variable
32
5 Conclusion
We provide evidence that investors pay attention to ESG ratings and use them in their
investment decisions. Using a quasi-natural experiment, of the adoption of Sustainalytics
new ESG Risk rating in October 2019 by Morningstar and Yahoo! Finance that changed
the ratings methodology and inverted the scale, we show that potentially incorrect assess-
ments about the meaning of the change in ESG ratings shape investors portfolio allocation
decisions and result in temporary price pressure on the a↵ected stocks.
We find that subsequent abnormal returns are positively related to the change in a
firm’s ESG rating. We show that firms with large negative changes in their ESG ratings
experience significantly negative abnormal returns while firms with positive changes have
positive abnormal returns. We interpret this as evidence that investors view the decline
or downgrade in ESG rating as bad news and the increase or upgrade as good news. We
show, however, that this interpretation is often not true. Many firms with declines in
their ESG rating have no change or are better ranked in terms of their ESG exposure
than their peers. Thus, some investors rely on ESG ratings when making an investment
decisions without adequate due diligence.
In order to understand the mechanism through which the implementation of the new
ESG rating methodology a↵ects prices of stocks, we investigate the responses of four types
of investors: individual investors, 13F institutions, ESG funds, and short-sellers. Using
33
34
28
Our sample is limited to U.S. firms and therefore, the e↵ect of the adoption of the change in method-
ology may be even greater than we document because it also a↵ects those foreign stocks that are rated
by Sustainalytics and available on Morningstar.
35
Angrist, J. D., Pischke, J.-S., 2008. Mostly harmless econometrics: An empiricist’s com-
panion. Princeton university press.
Ashcraft, A., 2010. MBS ratings and the mortgage credit boom. DIANE Publishing.
Barber, B. M., Morse, A., Yasuda, A., 2021. Impact investing. Journal of Financial
Economics 139, 162–185.
Barber, B. M., Odean, T., Zhu, N., 2006. Do Noise Traders Move Markets? Haas School
of Business Working paper pp. 1–47.
Berg, F., Fabisik, K., Sautner, Z., 2020. Rewriting history ii: The (un) predictable past
of esg ratings. European Corporate Governance Institute–Finance Working Paper 708.
Berg, F., Koelbel, J., Rigobon, R., 2019. Aggregate confusion: The divergence of ESG
ratings. Working Paper pp. 1–43.
Bialkowski, J., Starks, L. T., 2016. Sri funds: Investor demand, exogenous shocks and
esg profiles .
Biglaiser, G., 1993. Middlemen as experts. The RAND journal of Economics pp. 212–223.
Billio, M., Costola, M., Hristova, I., Latino, C., Pelizzon, L., 2020. Inside the ESG Rat-
ings: (Dis)agreement and performance. Corporate Social Responsibility and Environ-
mental Management Journal Forthcoming.
Boehmer, E., Huszar, Z. R., Jordan, B. D., 2010. The good news in short interest. Journal
of Financial Economics 96, 80–97.
Boehmer, E., Jones, C. M., Zhang, X., 2008. Unshackling short sellers: The repeal of the
uptick rule. Columbia Business School, unpublished manuscript, December .
Bolton, P., Freixas, X., Shapiro, J., 2012. The credit ratings game. The Journal of Finance
67, 85–111.
Bolton, P., Kacperczyk, M., 2021. Do investors care about carbon risk? Journal of
Financial Economics .
Calvet, L. E., Campbell, J. Y., Sodini, P., 2007. Down or out: Assessing the welfare costs
of household investment mistakes. Journal of Political Economy 115, 707–747.
Cao, J., Titman, S., Zhan, X., Zhang, W., 2020. Esg preference, institutional trading,
and stock return patterns. Tech. rep., National Bureau of Economic Research.
36
Célerier, C., Matray, A., 2019. Bank-branch supply, financial inclusion, and wealth accu-
mulation. The Review of Financial Studies 32, 4767–4809.
Chatterji, A. K., Durand, R., Levine, D. I., Touboul, S., 2016. Do ratings of firms con-
verge? Implications for managers, investors and strategy researchers. Strategic Man-
agement Journal 37, 1597–1614.
Chatterji, A. K., Levine, D. I., To↵el, M. W., 2009. How well do social ratings actu-
ally measure corporate social responsibility? Journal of Economics and Management
Strategy 18, 125–169.
Cohen, L., Diether, K. B., Malloy, C. J., 2007. Supply and demand shifts in the shorting
market. The Journal of Finance 62, 2061–2096.
Coval, J., Sta↵ord, E., 2007. Asset fire sales (and purchases) in equity markets. Journal
of Financial Economics 86, 479–512.
Delmas, M., Blass, V. D., 2010. Measuring corporate environmental performance: the
trade-o↵s of sustainability ratings. Business Strategy and the Environment 19, 245–
260.
Diether, K. B., Lee, K.-H., Werner, I. M., 2009. It’s SHO time! Short-sale price tests and
market quality. The Journal of Finance 64, 37–73.
Dorfleitner, G., Halbritter, G., Nguyen, M., 2015. Measuring the level and risk of cor-
porate responsibility - An empirical comparison of di↵erent ESG rating approaches.
Journal of Asset Management 16, 450–466.
Engelberg, J. E., Reed, A. V., Ringgenberg, M. C., 2012. How are shorts informed? Short
sellers, news, and information processing. Journal of Financial Economics 105, 260–278.
Fama, E. F., French, K. R., 1993. Common risk factors in the returns on stocks and
bonds. Journal of Financial Economics 33, 3–56.
Fama, E. F., French, K. R., 2007. Disagreement, tastes, and asset prices. Journal of
Financial Economics 83, 667–689.
Goldstein, I., Jiang, H., Ng, D. T., 2017. Investor flows and fragility in corporate bond
funds. Journal of Financial Economics 126, 592–613.
Goldstein, I., Kopytov, A., Shen, L., Xiang, H., 2021. On esg investing: Heterogeneous
preferences, information, and asset prices. Information, and Asset Prices (April 9, 2021)
.
37
Hirshleifer, D., 2001. Investor psychology and asset pricing. The journal of Finance 56,
1533–1597.
Hong, H., Kacperczyk, M., 2009. The price of sin: The e↵ects of social norms on markets.
Journal of financial economics 93, 15–36.
Hong, H., Li, F. W., Xu, J., 2019. Climate risks and market efficiency. Journal of econo-
metrics 208, 265–281.
Hsu, P.-H., Li, K., Tsou, C.-Y., 2020. The pollution premium. Available at SSRN 3578215
.
Ilhan, E., Sautner, Z., Vilkov, G., 2021. Carbon tail risk. The Review of Financial Studies
34, 1540–1571.
Kirchler, M., Huber, J., Stöckl, T., 2012. Thar she bursts: Reducing confusion reduces
bubbles. American Economic Review 102, 865–83.
Krueger, P., Sautner, Z., Starks, L. T., 2020. The importance of climate risks for institu-
tional investors. The Review of Financial Studies 33, 1067–1111.
Lizzeri, A., 1999. Information revelation and certification intermediaries. The RAND
Journal of Economics pp. 214–231.
Malenko, N., Shen, Y., 2016. The role of proxy advisory firms: Evidence from a regression-
discontinuity design. The Review of Financial Studies 29, 3394–3427.
Merton, R. C., 1987. A simple model of capital market equilibrium with incomplete
information. Journal of Finance 42, 483–510.
Mian, A., Sufi, A., 2011. House prices, home equity-based borrowing, and the US house-
hold leverage crisis. American Economic Review 101, 2132–2156.
Moss, A., Naughton, J. P., Wang, C., 2020. The irrelevance of esg disclosure to retail
investors: Evidence from robinhood. Available at SSRN 3604847 .
Pástor, L., Stambaugh, R. F., Taylor, L. A., 2020. Sustainable investing in equilibrium.
Journal of Financial Economics forthcoming.
Pedersen, L. H., Fitzgibbons, S., Pomorski, L., 2020. Responsible investing: The ESG-
efficient frontier. Journal of Financial Economics forthcoming.
38
Riedl, A., Smeets, P., 2017. Why do investors hold socially responsible mutual funds?
The Journal of Finance 72, 2505–2550.
Sangiorgi, F., Spatt, C. S., 2017. The economics of credit rating agencies. Foundations
and Trends in Finance 12, 1–116.
Semenova, N., Hassel, L. G., 2015. On the validity of environmental performance metrics.
Journal of Business Ethics 132, 249–258.
39
40
March 2016
Morningstar Sustain-
ability Rating launched September 2018
(from Sustainalytics) Sustainalytics introduces
new ESG Risk Rating
Figure 2: Comparison of Old ESG rating with New ESG Risk rating
This figure shows the distribution of the average old ESG rating in the three months (July-September
2019) before the adoption in Panel A and the average new ESG Risk rating in the three months (October-
December 2019) after the adoption of the new methodology by Morningstar and Yahoo! Finance in Panel
B.
Panel A
Panel B
42
43
Panel A
Panel B
44
Panel A
Panel B
45
46
AReti,t is a four-factor abnormal return estimated by using market beta from 12-month rolling-window
regression. ESG Pseudo Downgradei is an indicator variable equal to one if the change in ESGi
is in the lower quartile, and zero otherwise. We use October 2019 as month 0. di and dt denote stock
and year-month fixed e↵ects. The grey circles represent point estimates and the dash line is the 95%
confidence interval with standard errors clustered at the stock level.
47
Pre (Jul 2019 – Sep 2019) Post (Oct 2019 – Dec 2020)
Panel A: Levels Mean P50 SD P5 P95 Mean SD P50 P5 P95
ESG 50.51 48.00 7.23 42.00 65.00 30.38 30.02 10.36 14.59 49.05
1-F ARet (%) -0.90 -0.32 6.15 -12.09 7.77 0.10 -0.29 6.24 -8.76 10.37
3-F ARet (%) 0.41 0.68 6.20 -10.05 9.85 -0.02 -0.38 6.23 -9.09 10.23
4-F ARet (%) 0.45 0.65 6.26 -9.95 10.12 -0.08 -0.31 6.22 -9.08 10.17
# Retail (’00s) 27.50 2.91 153.20 0.18 75.67 29.20 3.09 164.63 0.18 78.73
% 13F Own 52.38 58.11 32.29 0.00 95.54 51.59 51.93 32.23 0.00 95.45
ESG Funds (%) 0.07 0.04 0.14 0.00 0.22 0.07 0.04 0.14 0.00 0.22
Short Int (%) 5.72 3.34 6.34 0.58 19.13 5.58 3.27 6.36 0.44 18.32
48
Sustainalytics Morningstar
Panel C: Correlations ESGi ESGi Ratingi ESG Ranki Classificationi
Electronic copy available at: https://ssrn.com/abstract=3801703
ESGi 1.0000
Panel A: Definitions
Old Morningstar Classification New Morningstar Classification
(Score relative to industry peers) (Risk relative to all firms)
50
in firm i’s ESG rating between the first month after the adoption (October 2019) and the last month (September 2019 ) before the adoption
of the new methodology by Morningstar and Yahoo! Finance. ESGi is the di↵erence between the average new ESG Risk rating in the three
months (October-December 2019) after the adoption and the average old ESG rating in the three months (July-September 2019) before the
adoption of the new methodology by Morningstar and Yahoo! Finance. ESG Pseudo-Downgradei is an indicator variable equal to one if
the change in ESGi is in the lower quartile, and zero otherwise. ESG Pseudo-Upgradei is an indicator variable equal to one if ESGi is
positive, zero otherwise. Postt is an indicator variable equal to one in the time period after the new ESG rating methodology is introduced,
zero otherwise. Sustainalytics Ratingi is the change in firm i’s Sustainanalytics average ESG Risk rating in the three months (October-
December 2019) after the adoption and the three months (July-September 2019) before the adoption of the new methodology by Morningstar
and Yahoo! Finance. ESG Ranki is the di↵erence in the firm’s relative ranking in the first month (October 2019) after the adoption and in
the last month (September 2019) before the adoption. Classification Upgradei (Classification Downgradei ) is an indicator variable
equal to one if the firm has an increase (decrease) in its Morningstar Classification after the adoption of the new methodology. Standard errors
51
are clustered at the firm level and t-statistics are reported in parentheses below the coefficient estimates. ***, **, and * denote significance at
the 1%, 5%, and 10%, respectively.
Observations 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914
R2 0.142 0.142 0.139 0.141 0.138 0.138 0.136 0.137 0.139 0.139 0.137 0.138
Table 3: ESG rating methodology change and abnormal returns (continued)
Observations 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914
R2 0.143 0.143 0.139 0.142 0.138 0.138 0.136 0.137 0.139 0.139 0.137 0.138
Table 3: ESG rating methodology change and abnormal returns (continued)
Firm FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year-Month FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914 12914
R2 0.143 0.143 0.140 0.142 0.138 0.138 0.137 0.138 0.139 0.139 0.138 0.139
Table 4: Change in abnormal returns for firms that have a decline in their ESG ratings and either no change or increase
in ranking
This table reports di↵erence-in-di↵erence regressions of the adoption of the new ESG Risk rating methodology by Morningstar on abnormal
returns. The sample is restricted to only those firms that have a decline in their ESG ratings and either no change or increase in ranking. The
dependent variables include in columns (1)-(4), a single-factor abnormal return estimated by using market beta from 12-month rolling-window
Electronic copy available at: https://ssrn.com/abstract=3801703
regression, in columns (5)-(8), three-factor abnormal returns and in columns (9)-(12), Carhart four-factor abnormal returns. Our sample includes
U.S. common stocks during two sub-periods: from July 2019 to September 2019 (three months the adoption) and October 2019 to December
2019 (three months after the adoption by Morningstar of the new methodology). ESGi is defined as the di↵erence in firm i’s ESG rating
between the first month after the adoption (October 2019) and the last month (September 2019 ) before the adoption of the new methodology by
Morningstar and Yahoo! Finance. ESGi is the di↵erence between the average new ESG Risk rating in the three months (October-December
2019) after the adoption and the average old ESG rating in the three months (July-September 2019) before the adoption of the new methodology
by Morningstar and Yahoo! Finance. ESG Pseudo-Downgradei is an indicator variable equal to one if the change in ESGi is in the lower
quartile, and zero otherwise. Postt is an indicator variable equal to one in the time period after the new ESG rating methodology is introduced,
zero otherwise. Standard errors are clustered at the firm level and t-statistics are reported in parentheses below the coefficient estimates. ***,
**, and * denote significance at the 1%, 5%, and 10%, respectively.
54
Observations 6493 6493 6493 6493 6493 6493 6493 6493 6493
R2 0.144 0.145 0.140 0.145 0.145 0.142 0.147 0.147 0.144
Table 5: ESG rating methodology change and change in ownership
This table reports OLS regressions of the adoption of the new ESG Risk rating methodology by Morningstar on the change in share ownership
or investor participation. The dependent variables are measured as the percentage change from the three month period (or quarter) after the
Electronic copy available at: https://ssrn.com/abstract=3801703
adoption to the three month period (or quarter) before the adoption of the new methodology by Morningstar and Yahoo! Finance. The dependent
variables include in columns (1)-(4), the change in retail participation measured as the change in the average number of Robinhood investors
holding a stock divided by the average dollar trading volume between April and June 2019, in columns (5)-(8), the quarterly change in holdings
by 13F institutions, in columns (9)-(12), the change in the holdings of funds identified as ESG-oriented funds, and in columns (13)-(15), the
change in short interest. Our sample includes U.S. common stocks during two sub-periods: from July 2019 to September 2019 (three months
before the adoption) and October 2019 to December 2019 (three months after the adoption by Morningstar of the new methodology). ESGi is
defined as the di↵erence in firm i’s ESG rating between the first month after the adoption (October 2019) and the last month (September 2019
) before the adoption of the new methodology by Morningstar and Yahoo! Finance. ESGi is the di↵erence between the average new ESG Risk
rating in the three months (October-December 2019) after the adoption and the average old ESG rating in the three months (July-September
2019) before the adoption of the new methodology by Morningstar and Yahoo! Finance. ESG Pseudo-Downgradei is an indicator variable
equal to one if the change in ESGi is in the lower quartile, and zero otherwise. ESG Pseudo-Upgradei is an indicator variable equal to one
55
if ESGi is positive, zero otherwise. Standard errors are adjusted for heteroscedasticity and t-statistics are reported in parentheses below the
coefficient estimates. ***, **, and * denote significance at the 1%, 5%, and 10%, respectively.
Observations 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310 2310
R2 0.010 0.010 0.004 0.005 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.007 0.007 0.002 0.006
Table 6: Change in ownership for firms that have a decline in their ESG ratings and either no change or increase in
ranking
This table reports OLS regressions of the adoption of the new ESG Risk rating methodology by Morningstar on the change in share ownership
or investor participation. The sample is restricted to only those firms that have a decline in their ESG ratings and either no change or increase
Electronic copy available at: https://ssrn.com/abstract=3801703
in ranking. The dependent variables are measured as the percentage change from the three month period (or quarter) after the adoption to the
three month period (or quarter) before the adoption of the new methodology by Morningstar and Yahoo! Finance. The dependent variables
include in columns (1)-(4), the change in retail participation measured as the change in the average number of Robinhood investors holding a
stock divided by the average dollar trading volume between April and June 2019, in columns (5)-(8), the quarterly change in holdings by 13F
institutions, in columns (9)-(12), the change in the holdings of funds identified as ESG-oriented funds, and in columns (13)-(15), the change in
short interest. Our sample includes U.S. common stocks during two sub-periods: from July 2019 to September 2019 (three months before the
adoption) and October 2019 to December 2019 (three months after the adoption by Morningstar of the new methodology). ESGi is defined
as the di↵erence in firm i’s ESG rating between the first month after the adoption (October 2019) and the last month (September 2019 ) before
the adoption of the new methodology by Morningstar and Yahoo! Finance. ESGi is the di↵erence between the average new ESG Risk rating
in the three months (October-December 2019) after the adoption and the average old ESG rating in the three months (July-September 2019)
before the adoption of the new methodology by Morningstar and Yahoo! Finance. ESG Pseudo-Downgradei is an indicator variable equal
to one if the change in ESGi is in the lower quartile, and zero otherwise. Standard errors are clustered at the firm level and t-statistics are
56
reported in parentheses below the coefficient estimates. ***, **, and * denote significance at the 1%, 5%, and 10%, respectively.
Observations 1190 1190 1190 1190 1190 1190 1190 1190 1190 1190 1190 1190
R2 0.016 0.016 0.004 0.003 0.003 0.002 0.001 0.001 0.000 0.007 0.007 0.003
Table 7: ESG rating methodology change and high institutional ownership
This table reports triple di↵erence-in-di↵erence regressions of the adoption of the new ESG Risk rating
methodology by Morningstar on abnormal returns for firms with high and firms with low institutional
ownership. High Insti is an indicator variable equal to one if the stock’s ownership is in both the
highest quartile of institutional ownership and the lowest quartile of retail participation, zero otherwise.
The dependent variables include in columns (1)-(2), a single-factor abnormal return estimated by using
market beta from 12-month rolling-window regression, in columns (3)-(4), three-factor abnormal returns
and in columns (5)-(6), Carhart four-factor abnormal returns. Our sample includes U.S. common stocks
during two sub-periods: from July 2019 to September 2019 (three months before the adoption) and
October 2019 to December 2019 (three months after the adoption by Morningstar and Yahoo! Finance
of the new methodology). ESGi is defined as the di↵erence in firm i’s ESG rating between the first
month after the adoption (October 2019) and the last month (September 2019 ) before the adoption
of the new methodology by Morningstar and Yahoo! Finance. ESGi is the di↵erence between the
average new ESG Risk rating in the three months (October-December 2019) after the adoption and
the average old ESG rating in the three months (July-September 2019) before the adoption of the new
methodology by Morningstar and Yahoo! Finance. Postt is an indicator variable equal to one in the
time period after the new ESG rating methodology is introduced, zero otherwise. ESG Ranki is
the di↵erence in the firm’s relative ranking in the first month (October 2019) after the adoption and
in the last month (September 2019) before the adoption. Sustainalytics Ratingi is the change in
firm i’s Sustainanalytics average ESG Risk rating the three months (October-December 2019) after the
adoption and in the three months (July-September 2019) before the adoption of the new methodology
by Morningstar and Yahoo! Finance. We include firm and year-month fixed e↵ects. Standard errors are
clustered at the firm level. t-statistics are reported in parentheses below the coefficient estimates. ***,
**, and * denote significance at the 1%, 5%, and 10%, respectively.
57
No. 307 Ester Faia, Andreas Fuster, Biases in Information Selection and
Vincenzo Pezone, Basit Zafar Processing: Survey Evidence from the
Pandemic
No. 305 Sabine Bernard, Benjamin Loos, The Disposition Effect in Boom and Bust
Martin Weber Markets
No. 304 Monica Billio, Andrew W. Lo, Global Realignment in Financial Market
Loriana Pelizzon, Mila Getmansky Dynamics: Evidence from ETF Networks
Sherman, Abalfazl Zareei
No. 302 Tim A. Kroencke, Maik Schmeling, The FOMC Risk Shift
Andreas Schrimpf
No. 301 Di Bu, Tobin Hanspal, Yin Liao, Risk Taking, Preferences, and Beliefs:
Yong Liu Evidence from Wuhan
No. 300 Dennis Gram, Pantelis An Extensible Model for Historical Financial
Karapanagiotis, Jan Krzyzanowski, Data with an Application to German Company
Marius Liebald, Uwe Walz and Stock Market Data