Journal Pre-proof

Do Analysts Matter for Green Investment? Evidence from the EU

Taxonomy

Grégoire Davrinche , Jean-Yves Filbien , Ludovic Vigneron

PII: S0165-1765(25)00114-4
DOI: https://doi.org/10.1016/j.econlet.2025.112277
Reference: ECOLET 112277

To appear in: Economics Letters

Received date: 9 December 2024

Revised date: 17 February 2025
Accepted date: 8 March 2025

Please cite this article as: Grégoire Davrinche , Jean-Yves Filbien , Ludovic Vigneron , Do Ana-
lysts Matter for Green Investment? Evidence from the EU Taxonomy, Economics Letters (2025), doi:
https://doi.org/10.1016/j.econlet.2025.112277

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2025 Published by Elsevier B.V.

 Highlights

 The study examines the impact of financial analysts' focus on environmental issues on corporate
green investments.
 Leverages the European Union Taxonomy Regulation for measuring corporate green
investments in French firms.
 Key finding: Analysts' focus is positively linked to green investments.

1
 Do Analysts Matter for Green Investment?

Evidence from the EU Taxonomy

Grégoire Davrinche, Jean-Yves Filbien, Ludovic Vigneron*

Abstract

To mitigate climate warming, one critical strategy is to redirect corporate investments toward low-carbon projects.
This study explores the role that financial analysts' focus on environmental issues can play in this process. The
research leverages the implementation of the EU Taxonomy Regulation for publicly listed companies in France
during the period 2022-2023, which provides a comprehensive, precise, and standardized measure of corporate
green investments. We use a firm-level measure of climate change exposure as a proxy for analysts’ attention on
environmental issues. We find that analyst attention has a positive and significant effect on corporate green
investments. Our results remain robust to potential self-selection bias and reverse causality. In a further analysis,
we also find that our main result is driven by analysts' focus on opportunities. This study offers new insights into
how analyst scrutiny impacts ESG practices and highlights the potential contributions of the EU Taxonomy
Regulation.

Keywords: Climate Change Exposure, Corporate Social Responsibility, ESG, EU Taxonomy Regulation,
Green Investments

* Grégoire Davrinche, TBS Business School, [email protected]; Jean-Yves Filbien

(corresponding author), Université Lille, LUMEN (ULR 4999), F-59000 Lille, France - 2 rue de Mulhouse 59000
Lille (France), [email protected]; Ludovic Vigneron, Université Lille, LUMEN (ULR 4999), F-
59000 Lille, France - 2 rue de Mulhouse 59000 Lille (France), [email protected]. Any errors or
omissions are the responsibility of the authors.

2
1. Introduction

Climate scientists predict rising global temperatures and an increase in their volatility, making severe weather
events nearly inevitable (Addoum et al., 2023). Growing awareness of the impact of economic activities on climate
change has prompted governments to accelerate the shift toward sustainability. As part of the European Green
Deal, the EU has committed to achieving carbon neutrality by 2050. On June 18, 2020, the EU adopted the
Taxonomy Regulation (TR) to “redirect capital flows towards sustainable investments and promote sustainable
and inclusive growth”. This regulation sets criteria for determining whether an economic activity is
environmentally sustainable.1 Since January 1, 2022, large European listed firms are required to report the
proportion of their CAPEX that is aligned with the TR.

The TR reflects the growing focus on ESG concerns among firms and their stakeholders. In recent decades, the
disclosure of information of a non-financial nature or relating to sustainability has progressively risen (Ferguson
et al., 2023). Prior literature documents a positive relationship between the disclosure of such information and firm
value (Buchanan et al., 2018). Hence, recent studies show that the mandatory reporting required by the TR conveys
value relevant information for market participants. Bassen et al. (2025) find a positive relationship between TR
alignment and stock returns, while Sautner et al. (2025) observe that firms with larger EU Taxonomy-eligible
revenue shares paid lower interest rates in the years before the formal introduction of the TR. Other studies show
that the KPIs mandated by Taxonomy reporting help reconcile and streamline the assessment of environmental
performance when ESG ratings diverge (Chrzan and Pott, 2024; Dumrose et al., 2022). Yet, there is limited
knowledge about the factors that might encourage firms to invest in greener projects, despite this being one of the
main objectives of the TR.

Numerous studies highlight the importance of attention in fostering the shift toward more sustainable practices.
Zhang et al. (2024) find that investor scrutiny pressures firms to improve their ESG standards. Wang and Chu
(2024) show that Chinese firms covered by ESG rating agencies significantly boost their green innovation. Another
research area links investor attention to market valuation, reflecting firms' financial decisions. Aouadi and Marsat
(2018) find that a strong CSR reputation positively impacts market value, but only for firms subject to high investor
scrutiny. He et al. (2024) find that media coverage effectively enhances corporate ESG performance. Among the
recent advancements in research on green finance, Sautner et al. (2023) constructed a new measure that captures
“the attention financial analysts and management devote to climate change topics at a given point in time” (Sautner
et al., 2023, P. 1450). This attention is proxied at the firm-level by the frequency of discussions relating to climate
change issues appearing in the transcripts of conference calls. In this paper, we investigate whether analyst
attention to climate change issues is correlated with firms’ CAPEX alignment to the EU TR.

Our empirical framework focuses on the TR because Europe is a pioneer in the field of extra-financial reporting.
Indeed, there is a strong regulatory push towards sustainability within the EU, which has implemented various
directives and regulations (see Online Appendix 1), such as the Non-Financial Reporting Directive (NFRD), and

1
To be considered aligned, an activity must: (i) contribute substantially to one or more of the environmental objectives, (ii) not
cause significant harm to any of the environmental objectives, and (iii) be carried out in compliance with minimum safeguards
(i.e. in compliance with international conventions on labor and human rights). At the core of the TR, the six environmental
objectives of the Taxonomy are: (1) climate change mitigation, (2) climate change adaptation, (3) sustainable use and protection
of water and marine resources, (4) transition to a circular economy, (5) pollution prevention and control, and (6) protection and
restoration of biodiversity and ecosystems.

3
more recently, the Corporate Sustainability Reporting Directive (CSRD) (which extends the scope of the NFRD).
Furthermore, European investors and consumers increasingly prioritize ESG considerations, creating market
demand for transparent and standardized extra-financial reporting (Amel-Zadeh and Serafeim, 2018; Krueger et
al., 2020). Finally, the EU TR offers firms a comprehensive, precise, and standardized approach to measuring
corporate green investment.2 This original framework contrasts with previous research, which have highlighted
significant inconsistencies and limitations in ESG ratings, casting doubt on their reliability in evaluating corporate
social and environmental performance (Berg et al., 2022; Yang, 2022).

We conducted our study on French firms listed on the CAC All-Tradable Index. Our study period starts in 2022,
the first year requiring full reporting on eligibility and alignment, and ends in 2023, the last year for which we
were able to collect data. We manually collected Taxonomy reporting data from Universal Registration Documents
(URDs), unlike prior studies that relied on databases like FTSE Russell's Green Revenues or ISS ESG (Dumrose
et al., 2022). This approach ensures our data accurately reflects the firms' reported indicators. Existing datasets
focus only on Taxonomy-aligned turnover, but we analyzed green CAPEX alignment to assess firms' commitment
to sustainable investment. Using the new measure of attention developed by Sautner et al. (2023), we found a
positive and significant correlation between analysts’ attention on climate change issues and green investment,
suggesting that greater analyst focus on climate change drives higher sustainable investment. Our results remain
robust after accounting for potential biases related to self-selection bias and reverse causality. In a further analysis,
we also find that financial analysts' focus on opportunities during conference calls drives increased green
investment. In contrast, the regulatory component is not correlated with green CAPEX, indicating that firms are
motivated by future economic benefits rather than compliance with existing regulations.

Our paper makes two key contributions: First, it extends existing research on the TR. While previous studies focus
mainly on the stock market effects of taxonomy-related disclosures (Bassen et al., 2025; Sautner et al., 2025) or
their relationship with traditional ESG ratings (Dumrose et al., 2022), the factors driving firms' commitment to
green activities remain underexplored. Unlike past work that emphasizes aligned turnover, our study focuses on
aligned CAPEX, which is forward-looking and better reflects a firm's long-term investment strategy and
commitment to green activities. Second, we contribute to the literature on attention and ESG performance. While
prior studies show a positive link between attention and ESG performance using proxies like Google Search
Volume Index (GSVI) or trading volume, we employ a novel measure developed by Sautner et al. (2023) that
captures market perceptions of a firm's exposure to climate-related factors.

2. Theoretical analysis and hypothesis development

To tackle climate change, production systems must transform significantly (Acemoglu et al., 2016). Firms need to
invest heavily in green initiatives but face challenges like technical hurdles, regulatory barriers, and funding
difficulties. One key issue is information asymmetry between management and investors (De Haas and Popov,
2023). Unlike traditional projects, green changes can render existing knowledge obsolete, deterring firms from
such investments. This exacerbates issues like adverse selection and moral hazard, leading to equity and debt
rationing (Xu and Kim, 2022). Additionally, CEO short-termism may hinder engagement in green projects due to

2The TR is a European regulation that applies immediately and mandatorily to the EU member states. Firms under the scope
of that regulation are legally required to provide a reporting, regardless of whether they meet the sustainability criteria or not.
This differs from "comply or explain" frameworks, where entities can opt out of compliance by providing a justification.

4
the time investors need to understand their value (Wang and Chu, 2024). Financial analysts play a crucial role in
mitigating these information asymmetries by gathering, reprocessing, and disseminating information, thereby
lowering the costs for investors and reducing information problems (Amiram et al., 2016; Chen et al., 2015). This
scrutiny can help address the incentive distortions between green and conventional projects.

Hypothesis 1: Increased analyst coverage of a firm is positively associated with its green investments.

To accurately assess green investments, analysts must develop skills beyond traditional finance, including
technical expertise and the ability to evaluate environmental performance. This shift aligns with the growing
investor preference for CSR investments (Bolton and Kacperczyk, 2021). ESG data firms now provide ratings like
financial ratings. Although ESG ratings from different providers vary (Berg et al., 2022; Gibson et al., 2021),
research shows they positively impact green investments (Gillan et al., 2021) and help reduce information
asymmetry on environmental performance. As climate change intensifies, firms are moving towards a low-carbon
economy. Analysts increasingly examine environmental issues to provide accurate assessments, a trend seen in
the rising focus on environmental topics in earnings calls (Hollander et al., 2010).

Hypothesis 2: Greater focus on environmental issues during a firm's earnings conference calls is positively
correlated with an increase in its green investments.

3. Method

3.1. Sample and data collection

Our initial sample is composed of French firms listed on the CAC All-Tradable index as of December 31, 2022
(239 firms), which we extracted from the Refinitiv database. The French setting is particularly relevant. First, it is
one of the leading economies in Europe, representing approximately 16.5% of the GDP in the EU. Second, France
has been at the forefront of environmental reporting in Europe for many years (New Economic Regulations in
2001, Grenelle Law in 2009-2010). Our study starts on the 2022 fiscal year, the first period in which firms were
required to furnish comprehensive data about TR. We exclude firms from the financial sector (ICB 1-digit), those
suspended from listing during the study period, and firms with fiscal years ending before December 31.3 We obtain
185 distinct firms for the 2022 fiscal year. We manually collected Taxonomy data directly from the reported
template presented in the URD in fiscal years 2022 and 2023. As depicted in Online Appendix 24, 123 firms report
the indicators stipulated by EU TR. We focus our analysis on the quantitative data from the mandatory templates.
For each economic activity, we collected CAPEX amounts (in euros) categorized as: (i) aligned with the taxonomy
(A1), eligible but not aligned (A2), and not eligible (B). For A1 and A2, firms must detail the amounts for each

3
Firms whose fiscal year straddles 2021 and 2022 (e.g. from September 1, 2021 to August 31, 2022) are only subject to
"eligibility" reporting. Excluding them allows us to obtain a sample of firms that are subject to equivalent obligations, ensuring
better comparability.
4 An initial analysis of the URDs was conducted to distinguish between firms that fall under the scope of the TR and those that

do not. To achieve that purpose, we performed an in-depth analysis of each URD to identify the presence of a section
specifically dedicated to extra-financial reporting. In France, firms subject to the TR are those required to produce a DPEF
(extra-financial performance document). This therefore implies, de facto, the obligation to provide information relating to the
TR. To ensure that we did not miss relevant information regarding Taxonomy disclosure outside a dedicated section, we also
conducted a keyword search on PDF files using the terms “taxonomy”. We also searched for the term “taxonimy” because
some firms mention it.

5
economic activity.5 CAPEX not eligible for EU TR (category B) can be reported in aggregate without detailing
specific economic activities.

3.2. Model

Our baseline regression is as follows:

Our dependent variable captures the actual firm level of green investment for firm i in year t compared to its
potential maximum level. Thus, AlignCapex represents the amount of CAPEX for eligible activities aligned with
the TR [A1 over (A1 + A2)].

To test our first hypothesis, we use analyst coverage (lnAnalystCoverage). We calculate this as the natural
logarithm of 1 plus the number of analysts providing at least one annual EPS forecast, based on Refinitiv data. For
the second hypothesis, we use Sautner et al.’s (2023) measure to proxy analysts’ attention to climate risks. They
assess firms' exposure to climate risk by analyzing the transcripts of earnings conference calls conducted with
financial analysts. They focus on how climate-related topics are raised and addressed during these interactive
segments. This approach is less prone to "greenwashing" than other corporate reports and has fewer selection bias
concerns, enhancing the reliability of the results. The algorithm uses bigrams to identify relevant sentences
discussing climate change, and the frequency of these bigrams is scaled by their total number in the transcript to
construct a broad measure of climate change exposure (CCExposure).6 Hence, their measure captures “the
attention financial analysts and management devote to climate change topics at a given point in time” (Sautner et
al., 2023, P. 1450).

To mitigate potential risks of reverse causality between our dependent variable and explanatory variables, we used
the lagged values of lnAnalystCoverage and CCExposure. We also include a set of control variables in the model.
EligCapex represents economic activities that are both aligned (category A1) and not aligned with the European
Taxonomy (category A2). Non-aligned activities are covered by the EU TR and could contribute to environmental
objectives. However, they do not meet the requirements defined in Article 8 of the EU TR. Our measure of
alignment allows us to account for the differences in eligibility level across firms in our sample. It captures the
firm’s effort regarding EU TR objectives. The rest of the control variables are financials and are defined in
Appendix 1. All continuous variables are winsorized at 1%. Finally, to account for potential correlated omitted
variables related to industry, we include fixed effects at the industry level (ICB 1-digit name). After integrating
analyst attention data and financial characteristics, our final sample comprises 79 distinct firms with complete data
for all variables for the two fiscal years 2022 and 2023 (representing a total of 158 firm-year observations).

5
Activities defined in the Commission Delegated Regulation (EU) 2021/2139 of 4 June 2021.
6
For a more detailed description of the variable construction, see Sautner et al. (2023).

6
4. Results and Discussion

4.1. Descriptive statistics

Table 1 shows the descriptive statistics for our sample. Firms report an average alignment of 29% with a median
close to 17%, and many report zero alignment. On average, 47% of CAPEX is eligible, with a median eligibility
of 45%, indicating less dispersion than for alignment. The mean difference analysis reveals that firms with high
alignment rates (above the median) have a significantly higher mean eligibility rate than those with low alignment.7
Collectively, these findings are consistent with existing evidence in the French context (PwC, 2023, AMF, 2023).

As mentioned earlier, we primarily use climate change exposure measures (multiplied by 10³ for clarity) derived
from Sautner et al. (2023). We find a difference in climate change exposure for firms in the high alignment group,
indicating higher scrutiny for firms in the high alignment group. We also collect the number of analysts covering
the stock and find no statistical differences between the two subsamples.

(Insert Table 1)

4.2. Main regression analysis

Previous literature often uses analyst coverage as a proxy for investor attention. In Column 1 of Table 2, we
observe that analyst coverage has no significant effect on the alignment rate using an OLS regression. This finding
does not corroborate our first hypothesis.8

(Insert Table 2)

In Column 2 of Table 2, CCExposure has a positive and highly significant coefficient (at the 1% level), indicating
that greater investor focus on climate change issues, as reflected in firm communications, correlates with increased
green investment. In Column 3 of Table 2, we present the results of our baseline equation presented in Section 3.2.
Of the variables related to investor attention, CCExposure remains significant at the 1% level, while
lnAnalystCoverage becomes positive and significant at the 10% level. Therefore, we emphasize the importance of
specific attention to CSR issues as a key determinant of green investment. Firms tend to report higher green
investments when they are closely scrutinized by financial markets on climate issues. This result is in line with
our second hypothesis.

Our descriptive statistics show that some firms report no aligned CAPEX, indicating left censoring. 9 In Column 4
of Table 2, we investigate the effect of the truncated dependent variable, first by using a Probit model.10 We find
that firms with fewer analysts covering them tend to have higher alignment rates (Wald chi2(18) = 33.51; Prob >
chi2 = 0.0145). Column 5 of Table 2 presents Tobit regression results, which confirm the positive and significant
coefficients (at the 1% level) of CCExposure. These last two results suggest that it is not the number of analysts

7
We also report the correlations among our main variables in Online Appendix 3.
8 We use the reghdfe package in
Stata, which automatically drops singleton observations. We obtain similar results if we include
them.
9 20.25% of observations found no aligned CAPEX.
10 In the fixed-effects logistic regression, only panels exhibiting variation in the outcome variable contribute meaningful

information. This accounts for the loss of observations.

7
following a firm that matters, but rather the intensity of focus on prominent green issues by a smaller group of
analysts.

4.3 The Heckman two-stage estimation

To account for potential self-selection bias, we implement a two-stage Heckman selection model (Heckman,
1979). A first stage logit regression of the firm's characteristics is used to quantify the probability of disclosing
EU TR templates based on financial characteristics and year / industry fixed effects. We also add the logarithm of
Total Carbon dioxide (CO2) and CO2 equivalents emission in tons to quantify the probability of disclosing EU
Taxonomy Regulation, since the technical expert group on sustainable finance (TEG) has determined the inclusion
in EU Taxonomy Regulation from CO2 emissions data. The estimation result of the first stage obtains the selection
bias correction term (inverse Mill’s ratio - IMR), which is introduced as an additional explanatory variable in the
second stage estimation. In Table 3 Columns 1 and 2, we report results of these two stages. We observe that the
coefficient associated with IMR is not significant, suggesting that our main results are not subject to a sample
selection issue.

(Insert Table 3)

4.4 Robustness tests

First, we test the possibility of reverse causality between our dependent and independent variables. This could
occur if AlignedCapex could explain our attention measures, making it difficult to determine the true direction of
our relationship. To properly test this possibility, we employ a Difference-in-Differences (DiD) approach, using
firms from the EU as the treated group and firms from the United Kingdom as the control group. The rationale is
that European regulations apply to EU firms, while the UK provides an empirical framework that closely resembles
that of the EU countries. The study period spans eight years, divided into a pre-taxonomy period (2016–2021) and
a post-taxonomy period (2022–2023). In the Online Appendix 4, we find no significant results associated with
postxtreated coefficient suggesting the lack of a reverse effect in our results. The corresponding results are
presented in the Online Appendix 4.11

Second, we also control for other attention proxies such as the GSVI (Da et al., 2011; Ferguson et al., 2023), the
occurrence of ESG controversies (Aouadi and Marsat, 2018) and the affiliation to a B2C sector (Flammer, 2015).12
As reported in Online Appendix 5, we find no significant results associated with the other attention proxies. We
argue that firms only respond to specific CSR issues highlighted by financial analysts.

4.5. Further analysis

11
We modify the definitions of the treated and control groups, and our main conclusion remains unchanged. Furthermore,
including control variables in a DiD analysis can be problematic, and they may be considered "bad controls" under certain
conditions. Thus, from Columns 5 and 8 of Online Appendix 4, we reproduce our analysis without including control variables.
Again, the lack of statistical significance for PostxTreated variable suggests that taxonomy reporting has not influenced analyst
attention.
12 We follow the same list of BtoC SIC codes as in Flammer (2015), which includes the following: 0000-0999, 2000-2399,

2500-2599, 2700-2799, 2830-2869, 3000-3219, 3420-3429, 3523, 3600-3669, 3700-3719, 3751, 3850-3879, 3880-3999, 4813,
4830-4899, 5000-5079, 5090-5099, 5130-5159, 5220-5999, 7000-7299, 7400-9999.

8
In an additional analysis, we aim to investigate which components of climate change exposure drive firms’
commitment to green investment. Sautner et al. (2023) develop three measures of climate change exposure:
opportunities (CCExposure_Op), regulations (CCExposure_Rg), and physical threats (CCExposure_Ph). In Table
4, we replace the CCExposure variable with each of these components in the baseline equation. It is noteworthy
that only the opportunities measure shows a significant (at 5% level) positive correlation with green CAPEX
investment, suggesting that financial analysts' focus on opportunities during conference calls drives increased
green investment.13 This result is consistent with the findings of Sautner et al. (2023), which show that the increase
in green-tech hiring, and green patents (crucial factors in the low-carbon transition) are mainly driven by the
opportunity component of their climate change measure. This suggests that innovative firms are more inclined to
invest in projects that are both more resilient and higher in risk. In contrast, the regulatory component is not
correlated with green CAPEX, indicating that firms are motivated by future economic benefits rather than
compliance with existing regulations. Finally, we find that the physical component of climate change exposure
(CCExposure_Ph) is not significantly correlated with green investment. This result is somewhat counterintuitive,
as firms exposed to physical climate risks might be expected to mitigate potential extreme weather impacts by
investing in more resilient activities. A possible explanation is that such investments do not necessarily align with
the sustainability criteria defined by the EU Taxonomy. For instance, relocating economic activities from
vulnerable areas is not inherently classified as sustainable under the Taxonomy framework.

(Insert Table 4)

Our findings highlight the significant role that intermediaries such as financial analysts can play in promoting
firms' green investments, in addition to the influence of regulatory frameworks. These results hold important
implications for all stakeholders seeking to exert pressure on firms, including shareholders and non-governmental
organizations (NGOs). Specifically, the findings suggest that stakeholders should leverage financial analysts as a
conduit for their demands, as this approach appears to be effective.

5. Conclusion

Climate scientists predict rising average temperatures and increased volatility, making severe weather events
almost inevitable. In response, governments are accelerating the shift toward sustainability, with the EU aiming
for carbon neutrality by 2050 as part of the European Green Deal. The EU TR, adopted on June 18, 2020, redirects
capital toward sustainable investments and sets criteria for determining the sustainability of economic activities.
It also requires large European firms to report their eligible CAPEX and aligned CAPEX with TR.

Our study focuses on French firms listed on the CAC All-Tradable Index in the two first years of full reporting on
eligibility and alignment. Using a firm-level climate change exposure metric from Sautner et al. (2023), we find a
positive and significant correlation between climate change exposure and green investment, suggesting that greater
attention paid to climate issues in firm communications leads to higher levels of green investment. In addition,
financial analysts' focus on opportunities during conference calls drives increased green investment. In contrast,
the regulatory component is not correlated with green CAPEX, indicating that firms are motivated by future
economic benefits rather than compliance with existing regulations.

13
We find similar results when using a Tobit model.

9
References

Acemoglu, D., Akcigit, U., Hanley, D., Kerr, W., 2016. Transition to clean technology. Journal of Political
Economy, 124(1), 52–104.
Addoum, J., Ng, D., Ortiz-Bobea, A., 2023. Temperature Shocks and Industry Earnings News. Journal of Financial
Economics, 150(1), 1-45.
Amel-Zadeh, A., Serafeim, G., 2018. Why and How Investors Use ESG Information: Evidence from a Global
Survey. Financial Analysts Journal, 74(3), 1‑17
Amiram, D., Owens, E., Rozenbaum, O., 2016. Do Information Releases Increase or Decrease Information
Asymmetry? New Evidence from Analyst Forecast Announcements. Journal of Accounting and Economics, 62(1),
121-138.
Aouadi, A., Marsat, S., 2018. Do ESG Controversies Matter for Firm Value? Evidence from International Data.
Journal of Business Ethics, 151(4), 1027-1047.
Bassen, A., Othar, K., Kerstin, L., Weiqiang, T., 2025. Revenue Alignment with the EU Taxonomy Regulation.
Journal of Banking & Finance, 170, 1-17.
Berg, F., Kölbel, J.F., Rigobon, R., 2022. Aggregate Confusion: The Divergence of ESG Ratings. Review of
Finance, 26(6), 1315–1344.
Bolton, P., Kacperczyk, M., 2021. Do investors care about carbon risk?. Journal of financial economics, 142(2),
517-549.
Buchanan, B., Cao, C., Chongyang, C., 2018. Corporate social responsibility, firm value, and influential
institutional ownership. Journal of Corporate Finance, 52, 73-95.
Chen, T., Harford, J., Lin, C., 2015. Do Analysts Matter for Governance? Evidence from Natural Experiments.
Journal of Financial Economics, 115(2), 383-410.
Chrzan, S., Pott, C, 2024. Limiting environmental reporting flexibility: investor judgment based on the EU
taxonomy. Review of Quantitative Finance and Accounting, 63(4), 1511-1548.
Da, Z., Engelberg, J., Gao, P., 2011. In Search of Attention. Journal of Finance, 66(5), 1461-1499.
De Haas, R., Popov, A., 2023. Finance and green growth. Economic Journal, 133(650), 637–668.
Dumrose, M., Rink, S., Eckert, J., 2022. Disaggregating confusion? The EU Taxonomy and its relation to ESG
rating. Finance Research Letters, 48(C), 102928.
Ferguson, P., Hronsky, J., Pinnuck, M., 2023. Who pays attention to sustainability reports and why? Evidence
from Google search activity. Accounting and Finance, 63(3), 3519-3551.
Flammer, C., 2015. Does product market competition foster corporate social responsibility? Evidence from trade
liberalization. Strategic Management Journal, 36, 1469–1485.
Gibson B.R., Krueger, P., Schmidt, P.S., 2021. ESG Rating Disagreement and Stock Returns. Financial Analysis
Journal, 77(4), 104-127.
Gillan, S.L., Koch, A., Starks, L.T., 2021. Firms and social responsibility: A review of ESG and CSR research in
corporate finance, 101889
He, F., Guo, X., Yue, P., 2024. Media coverage and corporate ESG performance: Evidence from China.
International Review of Financial Analysis, 91, 103003.
Heckman, J, 1979. Sample Selection Bias in Specification Error. Econometrica, 47(1), 153-161.

10
Hollander, S., Pronk, M., Roelofsen, E., 2010. Does Silence Speak? An Empirical Analysis of Disclosure Choices
During Conference Calls. Journal of Accounting Research, 48(3), 531-563.
Krueger, P., Sautner, Z., Starks, L. T., 2020. The Importance of Climate Risks for Institutional Investors. The
Review of Financial Studies, 33(3), 1067‑1111.
Sautner, Z., Van Lent, L., Vilkov, G., Zhang, R., 2023. Firm-Level Climate Change Exposure. The Journal of
Finance, 78(3), 1449-1498.
Sautner, Z., Yu, J., Zhong, R., Zhou, X., 2025. The EU Taxonomy and the Syndicated Loan Market. Journal of
Financial Services Research, forthcoming
Wang, Z., Chu, E., 2024. Shifting focus from end-of-pipe treatment to source control: ESG ratings’ impact on
corporate green innovation, Journal of Environmental Management, 354, 120409.
Xu, Q., Kim, T., 2022. Financial Constraints and Corporate Environmental Policies, The Review of Financial
Studies, 35(2), 576–635
Yang, R., 2022. What do we learn from ratings about corporate social responsibility? New evidence of
uninformative ratings. Journal of Financial Intermediation, 52, 100994.
Zhang, J., Zhang, L., Zhang, M., 2024. Media pressure, internal control, and corporate environmental information
disclosure. Finance Research Letters, 63, 105369.

11
 Table 1. Descriptive statistics

This table reports descriptive statistics and univariate tests related to the level of AlignCapex that is defined as a firm’s
substantial CAPEX contributions alignment divided by its CAPEX eligible to the EU Taxonomy. AnalystCoverage is the
number of analysts that make at least one annual EPS forecast. CCExposure measures the relative frequency with which climate
change bigrams occur in earnings calls following Sautner et al. (2023). EligCapex is defined as a firm’s CAPEX eligible to the
EU Taxonomy divided by its total CAPEX. Appendix 1 defines the remaining variables in detail. All financial continuous
variables are winsorized at 1%. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels respectively.

Mean Mean
Variables N Mean Median St-Dev. P25 P75 AlignCapex AlignCapex Mean Diff.
< Med. > Med.

EU Taxonomy reporting
AlignCapex 158 0,29 0,17 0,32 0,01 0,46 0 0,36 -0.358***
EligCapex 158 0,47 0,45 0,33 0,17 0,73 0,27 0,52 -0.245***
Investor attention
AnalystCoverage 158 15,1 15 7,25 9 21 13,78 15,44 -1,655
CCExposure_10_3 158 3,29 1,73 3,59 0,76 5,29 1,85 3,65 -1.801***
Financial data
Size 158 39,09 13,22 64,62 5,59 45,42 21,29 43,61 -22.314*
Profitability 158 0,07 0,06 0,08 0,04 0,11 0,05 0,07 -0,02
Leverage 158 0,28 0,28 0,14 0,17 0,38 0,24 0,29 -0,04
Investment 158 0,2 0,16 0,14 0,08 0,27 0,19 0,2 -0,007
Cash 158 0,14 0,13 0,09 0,08 0,17 0,18 0,13 0.045*
Dividend 158 0,85 1 0,36 1 1 0,78 0,87 -0,084

12
 Table 2. Main regression analysis

This table reports regressions that relate green investment and climate change investor attention. The dependent variable,
AlignCapex, is defined as a firm’s substantial CAPEX contributions alignment divided by its CAPEX eligible to the EU
Taxonomy Regulation. lnAnalystCoverage is the natural logarithm of 1 plus the number of analysts that make at least one
annual EPS forecast. CCExposure measures the relative frequency with which climate change bigrams occur in earnings calls
following Sautner et al. (2023). EligCapex is defined as a firm’s CAPEX eligible to the EU Taxonomy divided by its total
CAPEX. Appendix 1 defines the remaining variables in detail. All financial continuous variables are winsorized at 1%. ***,
**, and * indicate statistical significance at the 1%, 5%, and 10% levels respectively related to t-statistics that are robust to
heteroskedasticity. Standard errors are in parentheses.

(1) (2) (3) (4) (5)

lnAnalystCoveraget-1 0.079 0.115* -0.752 0.104

(0.068) (0.066) (0.555) (0.072)

CCExposuret-1 29.882*** 31.953*** 139.726** 37.605***

(9.182) (9.524) (66.795) (9.279)

EligCapext -0.108 -0.115 -0.110 1.888*** -0.055

(0.086) (0.083) (0.082) (0.565) (0.089)

lnSizet-1 0.011 0.019 -0.012 0.305 -0.004

(0.023) (0.015) (0.023) (0.197) (0.027)

Profitabilityt-1 -0.487 -0.290 -0.406 0.571 -0.375

(0.538) (0.566) (0.551) (1.972) (0.433)

Leveraget-1 -0.193 -0.058 -0.095 0.844 -0.031

(0.210) (0.216) (0.214) (1.384) (0.241)

Investmentt-1 0.569*** 0.302 0.396** -0.989 0.340

(0.187) (0.189) (0.190) (1.442) (0.245)

Casht-1 -0.072 -0.030 -0.101 -1.882 -0.171

(0.216) (0.222) (0.218) (1.692) (0.317)

Dividendt-1 -0.136 -0.132* -0.126 0.009 -0.134

(0.087) (0.076) (0.076) (0.484) (0.085)

Constant -0.054 -0.135 0.279 -4.420 0.123

(0.411) (0.355) (0.410) (3.562) (0.496)

Model OLS OLS OLS PROBIT TOBIT

IndustryFE Yes Yes Yes Yes Yes

YearFE Yes Yes Yes Yes Yes

N 158 158 158 150 158

Adj.R-Squared 0.359 0.413 0.427

AIC 162.17

13
 Table 3: The Heckman two-stage estimation
This table reports the results of a Heckman two-stage estimation of whether analyst attention affects green investment. In the
first stage (Columns 1), a probit model determines whether a firm disclosures EU Taxonomy Regulation template, with
explanatory financial variables including size, profitability, leverage, investment, cash, and dividend. We also add the logarithm
of Total Carbon dioxide (CO2) and CO2 equivalents emission in tons. We also include year fixed effects. After estimation, a
bias-correction term (IMR) for self-selection bias is obtained and becomes an additional explanatory variable for the second
stage. In the second stage estimation (Columns 2), explanatory variables include CCExposure, EligCapex, and the additional
selection bias correction term (IMR), and control variables that are the same as in regression equation. AlignCapex, is defined
as a firm’s substantial CAPEX contributions alignment divided by its CAPEX eligible to the EU Taxonomy. CCExposure
measures the relative frequency with which climate change bigrams occur in earnings calls following Sautner et al. (2023).
EligCapex is defined as a firm’s CAPEX eligible to the EU Taxonomy Regulation divided by its total CAPEX. TR Reporting
is a dummy variable equals to 1 if a firm provides EU Taxonomy Regulation template; 0 otherwise. IMR is the inverse Mills
ratio. Appendix 1 defines the remaining variables in detail. All financial continuous variables are winsorized at 1%. ***, **,
and * indicate statistical significance at the 1%, 5%, and 10% levels respectively related to t-statistics that are robust to
heteroskedasticity. Standard errors are in parentheses.

Dependent variables

TR Reporting AlignCapex

(1) (2)

CCExposuret-1 33.455***

(12.154)

IMR 0.229

(0.464)

EligCapext -0.197**

(0.098)

lnSizet-1 0.448*** 0.039*

(0.135) (0.023)

Profitabilityt-1 1.532 0.606

(2.866) (0.472)

Leveraget-1 1.152 0.066

(1.311) (0.299)

Investmentt-1 -1.747 0.381

(1.453) (0.234)

Casht-1 1.474 -0.308

(1.737) (0.247)

Dividendt-1 0.562 -0.109

(0.364) (0.104)

lnCO2t-1 1.745**

(0.819)

Constant -10.757*** -0.723

(3.232) (0.680)
Model PROBIT OLS

IndustryFE Yes Yes

YearFE Yes Yes

N 199 125

Adj. (pseudo) R-Squared 0.62 0.383

14
 Table 4. Main regression analysis by dimensions of climate change exposure
derived from conference call transcripts
This table reports regressions that relate green investment and climate change analyst attention. The dependent variable,
AlignCapex, is defined as a firm’s substantial CAPEX contributions alignment divided by its CAPEX eligible to the EU
Taxonomy RegulationlnAnalystCoverage is the natural logarithm of 1 plus the number of analysts that make at least one annual
EPS forecast. CCExposure_Op, CCExposure_Rg, and CCExposure_Ph, capture exposures respectively related to opportunity,
regulatory and physical shocks associated with climate change following Sautner et al. (2023). EligCapex is defined as a firm’s
CAPEX eligible to the EU Taxonomy divided by its total CAPEX multiplied by 100. Appendix 1 defines the remaining
variables in detail. All continuous variables are winsorized at 1%. ***, **, and * indicate statistical significance at the 1%, 5%,
and 10% levels respectively related to t-statistics that are robust to heteroskedasticity. Standard errors are in parentheses.

(1) (2) (3)

lnAnalystCoverage 0.110 0.081 0.086
(0.070) (0.068) (0.068)
CCExposure_Op 42.321**
(17.156)
CCExposure_Rg 14.164
(68.128)
CCExposure_Ph 422.101
(343.396)
EligCapex -0.109 -0.109 -0.093
(0.083) (0.087) (0.086)
Model OLS OLS OLS
Control Variables Yes Yes Yes
IndustryFE Yes Yes Yes
YearFE Yes Yes Yes
N 158 158 158
Adj.R-Squared 0.408 0.359 0.370

15
 Appendix 1. Variable definitions

Category Variable Definition Source

the firm’s substantial CAPEX contributions alignment divided Annual Report /
AlignCapex (%)
Taxonomy by its CAPEX exposed to the EU Taxonomy. Authors
Disclosure the firm’s CAPEX exposed to the EU Taxonomy divided by Annual Report /
EligCapex (%)
total firm's CAPEX. Authors
the natural logarithm of 1 plus the number of analysts that
lnAnalystCoverage (unit) Refinitiv
make at least one annual EPS forecast for a firm.
Sautner, Van Lent,
CCExposure the relative frequency with which climate change bigrams Vilkov, and Zhang
occur in earnings calls following Stauner et al. (2023) (2023)
the relative frequency with which bigrams that capture Sautner, Van Lent,
CCExposure_Op opportunities related to climate change occur in the transcripts Vilkov, and Zhang
of earnings conference calls following Stauner et al. (2023). (2023)
the relative frequency with which bigrams that capture
Sautner, Van Lent,
regulatory shocks related to climate change occur in the
CCExposure_Rg Vilkov, and Zhang
transcripts of earnings conference calls following Stauner et
(2023)
al. (2023).
the relative frequency with which bigrams that capture
Sautner, Van Lent,
physical shocks related to climate change occur in the
CCExposure_Ph Vilkov, and Zhang
Attention transcripts of earnings conference calls following Stauner et
(2023)
al. (2023).
the yearly average of GSVI that is the number of search query
GoogleSearchVolumeIndex submissions to the Google search engine on the related firm, Google Trends
(GSVI) (%) relative to the total number of query submissions on all kinds
of keywords which is then re-scaled on a range of 0 to 100.
a dummy variable equals 1 if the firm has at least one ESG
Controversies (dummy) Refinitiv
controversy between 2018 and 2022, 0 otherwise.
dummy variable equals to 1 when the firm belongs to a B2C
sector based on their 4-digit SIC codes, including the
following: 0000-0999, 2000-2399, 2500-2599, 2700-2799,
BtoC (dummy) 2830-2869, 3000-3219, 3420-3429, 3523, 3600-3669, 3700- Refinitiv
3719, 3751, 3850-3879,3880-3999, 4813, 4830-4899, 5000-
5079, 5090-5099, 5130-5159, 5220-5999, 7000-7299, 7400-
9999, 0 otherwise.
lnSize (euros) the natural logarithm of total assets in euros millions. Refinitiv
earnings before Interest and Taxes (EBIT) divided by total
Profitability (%)
assets.
Leverage (%) long-term debt divided by total assets. Refinitiv
Other firm Investment (%) Refinitiv
property, plant, and equipment (PPE) divided by total assets.
characteristics
Cash (%) cash holdings divided by total assets. Refinitiv
dummy variable equals to 1 if the firm pays a dividend per
Dividend (dummy) Refinitiv
share in 2021, 0 otherwise.
the natural logarithm of total Carbon dioxide (CO2) and CO2
lnCO2 (mill. tons) Refinitiv
equivalents emission in tons.

16