Climate Change Exposure and the Readability of Narrative Disclosures in Annual Reports

Abstract

This study investigates the influence of exposure to climate change on the readability of narrative disclosures in annual reports. Analyzing a sample of 38,229 firm-year observations from 2002 to 2022, the study provides evidence supporting the information obfuscation hypothesis. Specifically, it finds that exposure to climate change is linked to less readable annual reports. This effect is both statistically and economically significant; a one standard deviation increase in climate change exposure leads to an 8.5% reduction in readability. Moreover, this effect is particularly evident among firms operating in environmentally sensitive industries, as well as those characterized by weak corporate culture. Additional tests indicate that the different aspects of climate change exposure (opportunity, physical, and regulatory) are individually associated with a decrease in readability of annual reports, with the physical dimension exerting the most significant impact. The findings underscore the necessity of implementing measures to mitigate climate change exposure and enhance sustainable business environments, such as transitioning to renewable energy sources (such as solar, wind, and hydro), minimizing dependence on fossil fuels, minimizing emissions from industries and transportation, sourcing low-carbon materials, adopting circular economy models, directing capital toward climate-friendly projects, and managing climate risks through catastrophe bonds and climate insurance. The significance of these actions is underscored by the impact of climate change on firms’ information environments, as documented in the current study.

1. Introduction

Climate change significantly impacts firms’ operations and induces uncertainty surrounding their activities [1,2,3]. Extreme weather events brought on by climate change, for example, might interfere with company operations by altering how commodities and services are distributed and delivered [4,5,6]. Furthermore, extreme weather events may physically damage a company’s fixed assets, thereby diminishing its revenue-generating capacity [1]. These risks are further exacerbated by the lack of comprehensive climate risk insurance coverage for businesses. Given its substantial impact on companies’ operations and the uncertainty it generates, prior research has documented that climate change influences corporate decisions [7,8,9]. For example, studies have shown that climate change affects decisions related to cash holdings [10], financing choices [1], CEO compensation [11], corporate investment [12], leverage adjustment speed [13], supply chain management [14], and the allocation of resources to working capital [15].

Previous studies have largely focused on the impact of climate change on corporate decision-making, but little is known about how it affects the readability of annual reports. Understanding how climate change influences the readability of these reports is essential for developing a comprehensive understanding of how exposure to climate change influences corporate reporting decisions. Thus, this study aims to fill this void in the literature.

A significant portion of 10-K filings is dedicated to qualitative content, specifically narrative disclosures [16,17]. These disclosures aim to enhance market participants’ understanding of the decisions managers make when preparing and presenting financial data [18]. Readability is a crucial characteristic of narrative disclosures, as it enables individual investors and analysts to effectively interpret valuation-related information from financial statements [19,20]. It evaluates the ease with which the intended disclosures can be processed and understood [21,22]. Furthermore, readability has important implications for firm value [20], the cost of capital [23,24,25], equity pricing [26], access to trade credit [27], investment efficiency [28], and the risk of stock price crashes [29,30]. The destructive effect of complex disclosures on market participants’ ability to access written information underscores the importance of identifying the mechanisms that influence the readability of annual reports. Despite the critical role of readability in annual reports and the influence that managers have on the clarity of information by adjusting the readability of these reports, there is currently no evidence regarding the influence of climate change risk on the readability of annual reports.

The legitimacy theory posits that companies rely on society for resources and support to operate [31,32]. Consequently, they are compelled to act in accordance with accepted norms, values, and beliefs to legitimize their existence in the long term [31]. The risks associated with climate change affect companies’ compliance with social norms and values, creating a challenge for them to minimize the legitimacy gap [33]. To address this gap, companies may try to make their annual reports easily readable and understandable in case they are confronted with significant climate change risks to justify their exposure and manage market perceptions. This perspective suggests a potential enhancement of readability for firms with greater exposure to climate change.

However, the information obfuscation hypothesis suggests the presence of incentives for companies to conceal their poor performance by providing complex, lengthy, and less readable disclosures [16,34]. This obfuscation allows companies to defer or avoid unfavorable consequences in the capital markets. Supporting the obfuscation hypothesis, prior studies have documented evidence that companies obscure negative information in corporate reports through vague language [35,36,37]. Given the well-documented adverse impacts of climate change on how companies operate and perform [1,7,38], companies with significant exposure to climate change may downplay the influence of these hazards on their operations by offering complex, lengthy, and opaque disclosures. This perspective indicates a potential decrease in readability for firms with greater exposure to climate change. The aforementioned contradictory views highlight that the influence of climate change exposure on the readability of annual reports remains unclear.

The findings of the study suggest that climate change risk exposure is negatively related to the readability of annual reports. Furthermore, this effect is particularly pronounced in environmentally sensitive industries, where uncertainty related to climate exposure is significant compared to other sectors. We also find that the quality of corporate culture influences the link between climate change exposure and the readability of annual reports. Specifically, our results show that this negative impact of climate change exposure on readability is most evident in companies with a low-quality corporate culture.

In an additional test, we consider the possibility that companies with greater exposure to climate change may need to provide complex, detailed, and technical disclosures regarding their activities and initiatives. Consequently, measures of readability may indicate that reports from these companies are less comprehensible. While we cannot entirely dismiss the complexity explanation, we aim to further substantiate the obfuscation hypothesis. Specifically, we apply an alternative measure that directly reflects the transparency of annual reports: the natural logarithm of the word count in 10-K filings. If obfuscation motives are present, we expect climate change exposure to be associated with shorter annual reports, as this would indicate less transparency and a greater propensity to conceal poor performance. Conversely, if the complexity explanation prevails, we anticipate that firms with higher exposure will produce longer annual reports to elucidate the intricate effects of climate change. Our findings indicate that higher climate change risk is associated with shorter annual reports, despite the complexity linked to climate change exposure. This suggests that firms with greater exposure tend to produce more obscure annual reports, further supporting the obfuscation hypothesis.

We also examine whether the specific aspects of climate change exposure (opportunity, regulatory, and physical) lead to differences in the readability of annual reports across firms. The results of this disaggregation reveal that all three dimensions influence the readability of these reports, with the physical dimension exerting the most significant impact, which is due to the materiality of physical exposure to climate risk on firm outcomes [39,40].

This paper makes two important contributions to prior studies. The results of this study complement prior studies on factors affecting the readability of annual reports (e.g., [29,41,42,43]) by providing evidence on an important determinant that affects the readability of annual reports, which is climate change exposure. Furthermore, it adds to research examining the role that exposure to climate change plays in corporate decisions [12,14,15] by uncovering the influence of climate change risk on the readability of narrative disclosures. Moreover, this effect is confined to environmentally sensitive industries and companies with low cultural quality.

Furthermore, the results of the current study have significant implications due to the substantial economic effects associated with the low readability of annual reports [20,23,24,25]. Consequently, our results are crucial for regulators and market participants in promoting more transparent disclosures for companies affected by climate change, particularly those belonging to environmentally sensitive industries or possessing low-quality corporate cultures.

2. Hypotheses Development

The influence of exposure to climate change on the readability of annual reports remains unclear. Two mutually exclusive hypotheses could explain how climate change risk affects the readability of disclosures in 10-K reports: the legitimacy theory and the information obfuscation theory. The legitimacy theory relies on the premise that companies rely on society for resources and support to operate [31,32]. Consequently, companies are compelled to act in accordance with accepted norms, values, and beliefs; any violations of these principles can adversely affect a firm’s legitimacy, jeopardizing its existence [31]. Several actions taken by companies are motivated by the desire to build trust, influence societal perceptions, and legitimize the firm’s survival in the long term. For instance, Bansal and Roth [44] highlight that companies’ adoption of environmental initiatives is linked to their interactions with society and their efforts to obtain legitimacy by making their actions aligned with societal norms and values. Furthermore, companies often disclose private information when the market anticipates that they may withhold information in order to gain legitimacy and prevent their firm value from being discounted [45,46,47]. The risk of climate change represents a deviation from social norms and values, creating a challenge for companies to minimize the legitimacy gap [33]. Since legitimacy can be achieved if public perceptions are favorable, regardless of actual performance, companies may seek to minimize this gap by using annual reports as a legitimizing tool to explain their actions, shape public impressions, and mitigate social criticism. According to the legitimacy theory, companies may enhance the readability of their annual reports when exposed to high climate change risks to justify their exposure, discuss their actions, and manage market perceptions, thereby repairing, maintaining, or gaining legitimacy. Therefore, from the perspective of the legitimacy hypothesis, a positive association is expected between exposure to climate change and the readability of annual reports.

However, the information obfuscation hypothesis suggests that companies have incentives to obscure their poor performance by reducing the transparency of their information environment [16,34]. This obfuscation enables companies to avoid disclosing information that they prefer market participants not to be aware of, thereby deferring or avoiding unfavorable capital market consequences. One of the most common methods for reducing the transparency of the information environment is by producing hard to process narrative disclosures [16,34]. Prior studies have documented evidence that companies conceal negative information in corporate reports through vague and ambiguous language [35,36,37]. For instance, Li [16] found that companies issue less readable and longer annual reports when they experience low earnings and profitability. Lo et al. [17] discovered that companies tend to produce less readable narratives to amplify the upward management bias of their earnings to surpass the prior year’s results. Biddle et al. [28] showed that less readable disclosures are linked to investment inefficiency. Nguyen [48] finds that companies produce less readable disclosures to conceal their tax avoidance strategies. Given the detrimental effects of climate change on how companies operate and perform established in prior studies (e.g., [1,8,38]), companies with high exposure to climate change may obfuscate information to hide poor performance and obscure the influence of climate change on their operations by providing complex and opaque disclosures. This strategy aims to mitigate the negative market repercussions associated with greater exposure to climate risk. Therefore, from the perspective of the obfuscation hypothesis, a negative association is expected between exposure to climate change and the readability of annual reports.

Given these contradictory hypotheses, the following two competing hypotheses are considered:

Hypothesis 1a.

Climate change exposure is associated with a higher readability of annual reports.

Hypothesis 1b.

Climate change exposure is associated with a lower readability of annual reports.

We next consider that prior studies highlight that industry characteristics play an important role in firms’ response to climate change. Specifically, companies in specific industries—namely, those engaged in resource extraction and manufacturing—are particularly sensitive to environmental perspectives due to the direct impact of their operations on the environment, which may influence their response to climate change exposure. Prior studies have demonstrated that firms in environmentally sensitive industries exhibit different behaviors compared to those in less sensitive sectors, as market reactions to their environmental practices tend to differ from those of other firms [49,50]. Specifically, environmentally sensitive firms attract greater attention from investors and face intense pressure from stakeholders and social scrutiny due to the nature of their activities and the expectations placed upon them [44,49,51]. The theoretical and empirical implications of firms’ operations being directly linked to environmental resources have prompted studies to investigate the consequences of environmental sensitivity [52]. For instance, Bansal and Clelland [53] observe that environmentally sensitive firms encounter significant institutional pressure to improve their environmental performance, while Grougiou et al. [54] document negative public evaluations of these firms. Consequently, they are more concerned about the influence of their activities on the environment and bear a greater burden of environmental responsibility [55,56]. When considering their transparency, Appiagyei and Donkor [57] find that environmentally sensitive firms report less sustainability information with a regime of mandatory integrated reporting (IR). Accordingly, we posit that if climate change exposure results in a higher (lower) readability, the effect will be more evident for firms operating in highly environmentally sensitive industries, as these firms have stronger incentives related to their environmental impact. Hence, our next hypothesis is stated as follows:

Hypothesis 2.

The influence of climate change risk on the readability of annual reports is expected to be different for companies operating in environmentally sensitive industries.

Next, we consider how corporate culture influences the readability of firms’ annual reports [58], which may, in turn, affect the relation between climate change exposure and the readability of narrative disclosures. Firms with cultures characterized by strong quality norms tend to prioritize generating value through enhanced internal efficiency, refined processes, and a focus on long-term quality improvements [59]. As a result, climate change exposure may have a minimal impact on their operations, potentially influencing their disclosure incentives. In other words, despite their exposure to climate change, firms with strong quality culture may have little incentives to obscure information. Instead, they may have higher incentives to enhance the readability of their reports to differentiate themselves from to other firms and enhance their competitive advantages relative to other affected firms. Therefore, we hypothesize that if climate change exposure results in an increase (reduction) in readability, the effect is expected to be stronger (weaker) in the case of firms with a high-quality culture. Hence, our last hypothesis is stated as follows:

Hypothesis 3.

The effect of climate change risk on the readability of annual reports is expected to be different for in the case of firms with high quality culture.

3. Data and Methodology

3.1. Sample Selection

The sample constitutes U.S. companies in the Compustat between 2002 and 2022. We consider this period of time due to data availability on climate change exposure. We combine the Compustat dataset with the readability of annual reports as well as the variable capturing climate change exposure proposed by Sautner et al. [60]. We excluded observations on missing data on the readability, climate change exposure, or control variables yielded in 38,229 firm-year observations. These observations are used in the following analyses.

3.2. Methodology

We regress the readability of financial statements on climate change exposure to test our hypothesis:

$${{Readability}_{i,t}=\beta }_o+{\beta }_1 {Climate Exposure}_{i,t-1}+B_{2 }{\Phi }_{i,t-1}+B_{Year}+B_{Ind. }+{\epsilon }_{i,t}$$

(1)

where $Readability$ is of the proxy for the readability of annual reports, $Climate Exposure$ represents firm-level exposure to climate change risk, and $\mathsf{\Phi }$ represents a number of control variables. These covariates are defined in the following sections. Year ($B_{Year}$) and industry ($B_{Ind.}$) fixed effects are added to account for industry and year-specific effects that might drive our results. The subscripts i,t represent firm i at year t. We are interested in coefficient of $Climate Exposure$ in Equation (1).

3.2.1. Measuring Narrative Disclosures on Annual Reports ($Readability$)

Similarly to the wide range of recent literature [24,48,61], the bog index is used as the primary proxy of readability (we also utilize a different measure on the sensitivity tests). This variable was constructed by Bonsall et al. [61]. It provides a comprehensive assessment of plain English and overcomes the various problems associated with other measures of readability [48,62]. This index evaluates the various linguistic features of disclosures to assess the difficulty of reading 10-K reports, including weak verbs, passive voice, complex words, jargon, overused terms, and document length. Unlike traditional readability measures that rely on syllable counts or word length, the bog index refines readability scores by using a graded word list. In driving the values of the index, Bonsall et al. [61] utilized StyleWriter—The Plain English Editor, a computational linguistics software program to identify the attributes of plain English writing. In particular, the index is based on three components, using the following formula:

Bog Index = Sentence Bog + Word Bog − Pep

(2)

where ‘sentence bog’, captures sentence length issues and related difficulties, where longer sentences contribute to an increased bog index. The ‘Word bog’ assesses word complexity and stylistic problems in English by counting issues associated with these two components and applying a specific formula to evaluate the difficulty of reading words. The ‘Pep’ reflects how easily readers can comprehend and understand the text based on writing attributes by taking into account the use of good writing in the annual reports [61].

A lower bog index value indicates greater readability, while a higher value indicates lower readability [61]. Therefore, to simplify the interpretation, we multiply the index by −1 to derive the variable Readability.

3.2.2. Measuring Climate Change Risk

Climate risk is the variable proposed by Sautner et al. [60] for assessing firm-level exposure to climate change. This proxy has been extensively used in the literature [15,63,64]. Sautner et al. [60] account for the different costs incurred by firms owing to physical climate change impacts or regulatory measures aimed at mitigating global warming. Moreover, they consider that other firms simultaneously benefit from climate-related opportunities, e.g., renewable energy, electric vehicles, and energy storage. Additionally, they consider that companies within the same industry tend to experience different levels of climate risk over time, depending on their approach to sustainable policies (e.g., renewable resources and the net-zero transition). To account for these variations, Sautner et al. [33] developed a detailed measure that evaluates both climate change exposure and its impact across companies by analyzing the texts of earnings conference calls. Specifically, their measure is based on the proportion of conference call transcripts that relate specifically to climate change, reflecting the attention and concerns of market participants regarding its effects on individual firms. To ensure its reliability, Sautner et al. [60] conducted various validations, which confirmed that the measure effectively captures the time series and cross-sectional variations in climate change exposure.

3.2.3. Environmental Sensitivity

Following prior studies [65,66], we define environmentally sensitive firms based on their industries, where environmentally sensitive industries represent oil and gas, forestry, pulp and paper, energy, chemicals and drugs, mining and resources, and utilities (2-digit SIC codes, including 10, 12, 13, 14, 26, 28, 29, 33, 34, 36, 49).

3.2.4. Cultures with Strong Quality

We assess culture quality using the firm-specific text-based measure developed by Li et al. [59]. Specifically, Li et al. [59] provide a measure of corporate culture using the latest machine learning technique, which is a word-embedding model to analyze texts of earnings calls. Li et al. [59] conduct several tests to validate this measure of the quality of corporate culture.

3.2.5. Control Variables

$\mathsf{\Phi }$_i,t refers to a number of control variables related to annual reports (e.g., [17,29,41]). The control variables include the natural log of total assets (Size), market-to-book ratio (MB), profitability (ROA), return variability (Return variability), the count of special items in the annual reports (Special Items), the count of business segments (No. Bus. Segments) and the count of geographic segments the company has (No. Geo. Segments).

Table 1. Variable definitions.

Variable	Definition	Source
Readability	The bog index, which captures the readability of annual reports. In the analyses, we multiply the bog index by −1 to generate the readability measure.	Bonsall et al. [61]: https://www.dropbox.com/scl/fo/i9rcv7tnez343op1esvoy/AGfwWau87_mRf36JoKumRX0?rlkey=dqp0hwkcg6h0c0ofdji3wxw1e&e=1&st=o4lei4di&dl=0 (accessed on 15 December 2024)
Climate Exposure	Firm-level climate change exposure. We multiply this measure by 1000 to make the interpretation straightforward.	Sautner et al. [60]: https://osf.io/fd6jq/files/osfstorage (accessed on 15 December 2024)
Size	The natural log of total assets.	Compustat
ROA	Earnings before interest and taxes divided by total assets.	Compustat
MB	The market-to-book ratio.	Compustat
Return variability	The standard deviation of returns over the last five years.	Compustat
Special Items	The ratio of special items to total assets.	Compustat
No. Bus. Segments	The count of business segments	Compustat segment data
No. Geo. Segments	The count of geographic segments	Compustat segment data

The table defines the main variables and their sources.

presents the definitions and sources of the variables.

4. Results and Discussion

4.1. Summary Statistics

Table 2. Summary statistics.

	Count	sd.	Mean	p10	p25	p50	p75	p90
Bog Index	38,229	7.69	88.29	79.00	83.00	88.00	93.00	98.00
Climate Exposure	38,229	0.94	0.64	0.00	0.10	0.29	0.69	1.71
Size	38,229	1.91	6.85	4.38	5.57	6.83	8.09	9.32
ROA	38,229	0.19	0.03	−0.14	0.01	0.06	0.11	0.17
MB	38,229	1.55	1.80	0.65	0.90	1.29	2.07	3.51
Return variability	38,229	0.17	0.10	0.01	0.02	0.04	0.10	0.23
Special Items	38,229	0.06	−0.02	−0.05	−0.01	−0.00	0.00	0.00
No. Bus. Segments	38,229	2.10	1.95	0.00	0.00	2.00	4.00	5.00
No. Geo. Segments	38,229	2.42	2.58	0.00	0.00	2.00	4.00	6.00

The table displays the descriptive statistics. Covariates are winsorized at the 1st and 99th percentiles.

displays the descriptive statistics of the sample firms. The mean (median) bog index in our sample is 88.29 (88.00), and the mean (median) climate change exposure is 6.4% (2.9%). In the remaining tables, we use the variable Readability (which represents bog index multiplied by −1) to make the interpretation more straightforward. Moreover, the descriptive statistics show a significant variation in the sample firms’ climate change exposure, which highlights the importance of examining the effect of such variability. Furthermore, the descriptive statistics show the existence of variations in sample size, which suggests that our sample comprises firms of different sizes.

Table 3. Correlation matrix.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
(1) Readability	1.00
(2) Climate Exposure	−0.07 ***	1.00
(3) Size	0.11 ***	0.02 **	1.00
(4) ROA	−0.19 ***	−0.01	0.39 ***	1.00
(5) MB	0.09 ***	−0.06 ***	−0.17 ***	−0.04 ***	1.00
(6) Return variability	0.07 ***	−0.02 **	−0.40 ***	−0.50 ***	0.19 ***	1.00
(7) Special Items	−0.03 ***	0.01 *	0.11 ***	0.17 ***	0.06 ***	−0.23 ***	1.00
(8) No. Bus. Segments	0.04 ***	0.13 ***	0.34 ***	0.17 ***	−0.20 ***	−0.17 ***	−0.01	1.00
(9) No. Geo. Segments	0.05 ***	0.15 ***	0.08 ***	0.07 ***	0.03 ***	−0.03 ***	−0.05 ***	0.17 ***	1.00

The table displays the correlation matrix. All covariates are defined in Table 1. ***, ** and * represent 1%, 5%, and 10% significance, respectively.

displays the correlation matrix. It highlights that climate change exposure negatively and significantly correlated with the readability of annual reports. As shown in Column 1, the correlation coefficient between Climate Exposure and Readability is −7% (p < 0.01). This initially supports the information obfuscation hypothesis. Moreover, there are no significant correlations between the independent variables as they are less than the critical value of 0.70, indicating no multicollinearity issues.

4.2. Empirical Results

Table 4. Main results.

	(1)	(2)
Climate Exposure	−0.06 * (−1.69)	−0.09 ** (−2.26)
Size		0.56 *** (25.40)
ROA		−4.70 *** (−19.82)
MB		−0.05 ** (−2.16)
Return variability		1.44 *** (5.43)
Special Items		−2.30 *** (−3.97)
No. Bus. Segments		0.13 *** (7.76)
No. Geo. Segments		−0.03 * (−1.75)
Constant	88.33 *** (2336.88)	84.39 *** (510.01)
Year FE	Yes	Yes
Industry FE	Yes	Yes
Observations	38,229	38,229
Adjusted R2	0.411	0.432

The table displays the regression analyses for examining Hypothesis 1 on the role of climate change exposure on the readability of annual reports. Column 1 presents a basic specification that regresses readability on climate change exposure without control variables, while Column 2 presents the regression analysis for the main analyses as appear in Equation (1). All continuous variables are winsorized at the 1st and 99th percentiles. The t-statistics reported are based on firm clusters and heteroskedasticity-corrected standard errors. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

displays the results of Hypotheses 1a and 1b regarding the influence of exposure to climate change risk on the readability of annual reports, as outlined in Equation (1). Column 1 presents a basic model that regresses Readability on Climate Exposure, without including additional covariates. The simple model indicates that climate change exposure is negatively associated with the readability. This provides initial evidence that climate change risk adversely affects the readability of annual reports.

Column 2 displays the main specification while also incorporating additional variables, as illustrated in Equation (1). The results indicate that climate change risk is negatively and significantly related to the readability of narrative disclosures in annual reports (based on the variable Readability, which is bog index*-1). Specifically, the coefficient for climate change risk (Climate Exposure) is −0.09 (p < 0.05). This influence is also economically significant. In particular, the evaluation of economic significance reveals that an increase of one standard deviation in Climate Exposure leads to an 8.5% reduction in the readability of annual reports −0.09 * 0.94, calculated based on the standard deviation of Climate Exposure reported in Table 2. This finding supports Hypothesis 1b, which predicts a negative relation between climate change exposure and the readability of annual reports. The control variables align with expectations and are consistent with those of prior studies.

These results illuminate the role of climate change exposure in firms’ information environment by incentivizing them to produce less-readable annual reports. This finding aligns with the information obfuscation hypothesis, which highlights a tendency for firms to obscure unfavorable news in order to conceal poor performance and diminish the perceived impact of climate change on their operations, thereby avoiding negative market scrutiny regarding their exposure. These findings add to prior studies on firms’ tendency to produce hard to process narrative disclosures or vague and ambiguous language to hide bad information [16,34,35,36,37]. Our findings emphasize that the effects of climate change exposure may extend beyond firm performance, as they also appear to influence the reporting decisions made by managers in the current study.

Table 5. Environmental sensitivity.

	(1)	(2)
	High Enviro. Exposure	Low Enviro. Exposure
Climate Exposure	−0.23 *** (−3.93)	0.04 (0.74)
Size	0.52 *** (11.40)	0.56 *** (22.33)
ROA	−5.28 *** (−14.00)	−3.45 *** (−11.27)
MB	0.28 *** (5.85)	−0.20 *** (−7.32)
Return variability	0.85 * (1.85)	1.48 *** (4.70)
Special Items	−2.43 ** (−2.11)	−2.70 *** (−4.08)
No. Bus. Segments	−0.01 (−0.17)	0.18 *** (9.75)
No. Geo. Segments	−0.17 *** (−6.01)	0.07 *** (3.80)
Constant	86.54 *** (256.84)	83.60 *** (443.07)
Year FE	Yes	Yes
Industry FE	Yes	Yes
Observations	10,193	28,036
Adjusted R2	0.428	0.426

The table displays the regression analyses for examining Hypothesis 2 on the effect of environmental sensitivity. Column 1 displays the findings for the subsample of companies in industries with high environmental sensitivity, while Column 2 displays the findings for companies in industries with low environmental sensitivity. The t-statistics reported are based on firm clusters and heteroskedasticity-corrected standard errors. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

displays the regression analyses for testing Hypothesis 2, which examines the effect of industry sensitivity to the environment regarding how climate change exposure influences the readability of annual reports. Column 1 displays the findings for the subsample of companies in industries with high environmental sensitivity, while Column 2 displays the findings for companies in industries with low environmental sensitivity. The findings indicate that the influence of climate risk on the readability of annual reports is evident mainly among firms in environmentally sensitive industries. Specifically, the coefficient for climate change risk (Climate Exposure) is −0.23 (p < 0.01) in Column 1. In contrast, the coefficient for climate change risk (Climate Exposure) in Column 2 is 0.04, which is statistically insignificant.

These results support Hypothesis 2, suggesting that companies’ responses to climate change risk are influenced by their industries’ sensitivity to environmental factors. Given their high reliance on environmental resources, firms in environmentally sensitive industries seem to obscure the detrimental effects of climate change on their operations by producing less-readable annual reports. This finding adds to prior studies on the influence of industry characteristics, particularly sensitivity to environmental aspects on firms’ response to climate change [54,55,56,57,67]. Our findings indicate that this influence of industry extends to the effect of climate change on the readability of annual reports.

Table 6. Quality culture.

	(1)	(2)
	High-quality culture	Low-quality culture
Climate Exposure	0.06 (1.09)	−0.14 ** (−2.42)
Size	0.68 *** (18.46)	0.40 *** (12.91)
ROA	−3.73 *** (−9.70)	−5.81 *** (−16.88)
MB	−0.18 *** (−4.65)	0.03 (1.00)
Return variability	1.48 *** (3.63)	1.28 *** (2.96)
Special Items	−2.69 *** (−2.97)	−2.65 *** (−3.08)
No. Bus. Segments	0.11 *** (4.17)	0.14 *** (6.14)
No. Geo. Segments	0.03 (1.18)	−0.06 *** (−2.85)
Constant	83.76 *** (310.84)	85.16 *** (359.29)
Year FE	Yes	Yes
Industry FE	Yes	Yes
Observations	13,547	19,772
Adjusted R2	0.362	0.473

The table displays the regression analyses for examining Hypothesis 3 on the influence of quality culture. Column 1 displays the findings for the subsample of firms with a high-quality culture, while Column 2 displays the findings for the subsample of firms with a low-quality culture. The t-statistics reported are based on firm clusters and heteroskedasticity-corrected standard errors. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

displays the regression analyses for testing Hypothesis 3, considering the impact of quality culture within firms on how climate change exposure influences the readability of annual reports. Column 1 displays the findings for the subsample of firms with a high-quality culture, while Column 2 displays the findings for the subsample of firms with a low-quality culture. The findings indicate that the influences of climate risk on the readability of annual reports are evident only among firms characterized by cultures of low-quality norms. Specifically, the corresponding coefficient for climate change risk (Climate Exposure) is −0.23 (p < 0.01), as shown in Column 1. However, the corresponding coefficient for climate change risk (Climate Exposure) is 0.04 and statistically insignificant in Column 2. These results support Hypothesis 3, suggesting that corporate culture influences companies’ responses to climate change risk. Firms with a culture characterized by low-quality norms tend to obscure the detrimental impact of climate change on their operations by producing less readable annual reports. This finding complements prior studies on the influence of corporate culture on the readability of annual reports [58]. Our findings indicate that the existence of a strong culture mitigates the effect of climate change on the readability of annual reports, as only firms with a weak culture respond to climate change risk by reducing the readability of their annual reports.

Overall, the regressions presented in Table 4, Table 5 and Table 6 show that the information obfuscation hypothesis predominates over the legitimacy hypothesis when investigating the influence of climate change exposure on companies reporting decisions. In particular, climate change exposure seems to motivate firms to produce less readable narrative disclosures in order to obscure the detrimental influence of climate change. Moreover, this effect is particularly evident among companies operating in environmentally sensitive industries, where uncertainty is exacerbated by sector-specific factors. Furthermore, strong quality cultures mitigate firms’ tendency to obscure information, as the documented effect is observed only in firms with low-quality cultures.

5. Alternative Effect of Climate Change Exposure

In the following tests, we investigate the possibility that firms with greater exposure to climate change may need to provide complex, detailed, and technical disclosures regarding their activities and initiatives. As a result, both machines and humans may find these reports less readable. While we cannot completely dismiss the complexity explanation, we aim to further substantiate the obfuscation hypothesis. To address this concern, we employ an alternative measure that directly reflects the transparency of annual reports. Specifically, following Hasan [62], we utilize the natural logarithm of the word count in 10-K filings. This measure serves as a direct proxy for the readability and transparency of financial disclosures [19]. It reflects the level of detail in the information provided by firms in their annual reports, a distinct aspect of firms’ clarity in delivering narrative disclosures. Therefore, if obfuscation motives are present (i.e., firms attempt to conceal poor information), we expect climate change exposure to be associated with shorter annual reports. This would undermine the complexity explanation of the findings. Conversely, if the complexity explanation holds true (i.e., the lower readability is due to firms providing detailed and technical disclosures regarding their activities and initiatives related to climate change), we anticipate that firms with higher exposure will produce longer annual reports to clarify the intricate effects of climate change. We obtained the data from the Loughran and McDonald website (available at: https://sraf.nd.edu/sec-edgar-data/lm_10x_summaries/, accessed on 10 May 2025). The results are displayed in

Table 7. Alternative effect of climate change exposure.

	(1)	(2)	(3)	(4)	(5)
	Full sample	High Enviro. Exposure	Low Enviro. Exposure	High-quality culture	Low-quality culture
Climate Exposure	−2.23 *** (−2.88)	−1.44 ** (2.38)	−5.43 (−1.56)	6.90 (1.56)	−9.77 ** (−2.43)
Size	0.11 *** (73.72)	0.11 *** (35.45)	0.11 *** (64.58)	0.12 *** (48.30)	0.10 *** (46.32)
ROA	−0.42 *** (−29.30)	−0.38 *** (−17.20)	−0.45 *** (−22.92)	−0.44 *** (−18.86)	−0.40 *** (−19.17)
MB	−0.00 ** (−2.19)	−0.01 ** (−2.25)	−0.00 (−0.82)	−0.00 (−0.13)	−0.00 ** (−2.04)
Return variability	0.23 *** (15.27)	0.21 *** (9.38)	0.24 *** (11.61)	0.23 *** (9.00)	0.25 *** (10.74)
Special Items	−0.35 *** (−8.71)	−0.42 *** (−5.78)	−0.32 *** (−6.68)	−0.35 *** (−5.23)	−0.42 *** (−7.52)
No. Bus. Segments	0.01 *** (6.07)	0.00 (0.36)	0.01 *** (7.03)	0.01 *** (4.89)	0.01 *** (4.41)
No. Geo. Segments	0.00 ** (2.22)	−0.00 (−1.01)	0.00 *** (3.76)	0.00 (0.97)	0.00 * (1.81)
Constant	10.09 *** (938.63)	10.16 *** (474.46)	10.06 *** (806.99)	10.00 *** (588.89)	10.17 *** (637.57)
Year FE	Yes	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes	Yes
Observations	36,985	9869	27,116	13,160	19,115
Adjusted R2	0.273	0.263	0.275	0.279	0.262

The table presents the results for using the length of annual reports. Column 1 presents the results for Hypothesis 1, Columns 2 and 3 present the results for Hypothesis 2, while Columns 4 and 5 present the results for Hypothesis 3. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

.

We find that a higher risk of climate change is associated with shorter annual reports, despite the complexities related to climate change exposure. This suggests that firms with greater exposure tend to produce more ambiguous annual reports, further supporting the obfuscation hypothesis.

6. Additional Test and Robustness Checks

6.1. Different Aspects of Climate Change Exposure

We consider whether the documented effect is specific to a certain dimension of exposure to climate change risk. Specifically, we consider the different climate change shocks: opportunity shock, physical shock, and regulatory shocks based on the disaggregated climate change “topics” measures proposed by Sautner et al. [60]. We run the model to estimate climate change exposure for different topics separately and present the results in

Table 8. Different aspects of climate change exposure.

	(1)	(2)	(3)
Opport. Exposure	−0.21 * (−1.74)
Regulatory Exposure		−1.10 * (−1.78)
Physical Exposure			−5.29 *** (−3.04)
Size	0.56 *** (25.45)	0.56 *** (25.54)	0.56 *** (25.53)
ROA	−4.70 *** (−19.82)	−4.69 *** (−19.79)	−4.70 *** (−19.81)
MB	−0.05 ** (−2.14)	−0.05 ** (−2.14)	−0.05 ** (−2.10)
Return variability	1.44 *** (5.43)	1.44 *** (5.45)	1.44 *** (5.42)
Special Items	−2.31 *** (−4.00)	−2.32 *** (−4.01)	−2.30 *** (−3.98)
No. Bus. Segments	0.13 *** (7.69)	0.12 *** (7.65)	0.12 *** (7.66)
No. Geo. Segments	−0.03 * (−1.78)	−0.03 * (−1.83)	−0.03 * (−1.90)
Constant	84.36 *** (513.33)	84.34 *** (515.88)	84.35 *** (515.99)
Year FE	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes
Observations	38,229	38,229	38,229
Adjusted R2	0.432	0.432	0.432

The table displays the results for testing the specific aspects of climate change exposure. Column 1 displays the regression for opportunity exposure, Column 2 displays the regression for regulatory exposure, and Column 3 displays the regression for physical exposure. The t-statistics reported are based on firm clusters and heteroskedasticity-corrected standard errors. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

. Column 1 displays the regression for opportunity exposure, Column 2 displays the regression for regulatory exposure, and Column 3 displays the regression for physical exposure.

The results presented in the table indicate that the three dimensions of climate change exposure negatively affect the readability of annual reports. The corresponding coefficient of Opport. Exposure is −0.21 (p < 0.1), the coefficient of Regulatory Exposure is −1.10 (p < 0.1), and the coefficient of Physical Exposure is −5.29 (p < 0.01). Yet, we note that physical exposure has the strongest effect on the readability of annual reports. This could be illuminated by the difficulty in anticipating physical exposure to climate risk which leads its effect to be material [39,40]. Hence, firms seem to respond more strongly to this form of exposure.

6.2. Robustness Checks

We conduct different sensitivity tests to confirm the robustness of our results. We summarize them in

Table 9. Sensitivity tests.

	(1)	(2)	(3)	(4)	(5)
	Full sample	High Enviro. Exposure	Low Enviro. Exposure	High-quality culture	Low-quality culture
Panel A: Time series changes in public consideration of climate risk
Climate Exposure	−0.08 ** (−2.20)	−0.23 *** (−3.89)	0.05 (1.08)	0.07 (1.19)	−0.15 ** (−2.36)
Panel B: Additional firm-level variables
Climate Exposure	−0.09 ** (−2.31)	−0.23 *** (−4.07)	0.04 (0.82)	0.08 (1.46)	−0.14 *** (−2.74)
Panel C: Excluding years of high uncertainty
Climate Exposure	−0.08 * (−1.86)	−0.22 *** (−3.40)	0.04 (0.69)	0.09 (1.55)	−0.14 ** (−2.10)
Panel D: State fixed effects
Climate Exposure	−0.07 ** (−2.19)	−0.22 *** (−2.60)	0.07 (1.44)	0.13 (1.19)	−0.11 ** (−2.12)
Panel E: Adding financial firms
Climate Exposure	−0.08 ** (−2.07)	−0.23 *** (−3.93)	0.05 (0.96)	0.06 (1.12)	−0.15 ** (−2.33)

The table displays the robustness tests of Hypotheses 1, 2, and 3. Column 1 presents the robustness tests for Hypothesis 1, Columns 2 and 3 present the robustness tests for Hypothesis 2, while Columns 4 and 5 present the robustness tests for Hypothesis 3. The t-statistics reported are based on firm clusters and heteroskedasticity-corrected standard errors. ***, **, and * represent 1%, 5%, and 10% significance, respectively.

(for the sake of brevity, we only show the results for the main variable). First, we assess the robustness of our findings by accounting for overall time series fluctuations in public consideration/awareness of climate change risk. Specifically, we rely on the index proposed by Engle et al. [68]. This measure proxies for the amount of climate-related news in the Wall Street Journal. Panel A of Table 9 shows that our results are robust to account for overall time series fluctuations in public awareness of climate change risk. Specifically, the variable Climate Exposure is negative and significantly related to the readability of annual reports.

Second, additional firm-level variables, such as industry average return variability and leverage, and the natural log of company age were added to confirm that the results are robust to a different model specification. Panel B of Table 9 shows that our inferences are similar when adding additional firm-level variables.

Third, firm-year observations from 2008, 2009, and 2020 were excluded to eliminate the effect of the financial crisis and COVID-19 pandemic on the findings. Panel C of Table 9 shows that our inferences are not affected by the exclusion of these periods.

Fourth, state-fixed effects were added to eliminate the effect of variations across companies attributed to location. Panel D of Table 9 displays that our inferences are similar when controlling for state-fixed effects.

Finally, financial firms were excluded to eliminate the effect of regulations and different business structure for these firms on the potential bias of the measure of readability. Panel E of Table 9 shows that our inferences are qualitatively similar.

Taking these together, the results of different tests indicate that our inferences are robust.

7. Conclusions

This study demonstrates that companies with a high exposure to climate change produce less-readable annual reports. This effect is particularly pronounced among companies operating in environmentally sensitive industries and those with low-quality corporate cultures. Furthermore, the results of disaggregating the measure of climate change exposure into its subdimensions indicate that all three dimensions influence the readability of these reports, with the physical dimension exerting the most significant impact.

We build upon previous research regarding the factors influencing the readability of annual reports. This research indicates that elements such as corporate performance, investment efficiency, earnings management, and tax avoidance strategies significantly impact the readability of annual reports. Additionally, we expand on studies examining the consequences of climate change risk exposure, which demonstrate its effects on profitability, access to capital, and corporate decision-making, among other outcomes.

The findings from this study have significant implications due to the substantial economic effects resulting from the low readability of annual reports. Consequently, our results are crucial for regulators and market participants in promoting more transparent disclosures for firms affected by climate change, particularly those operating in environmentally sensitive industries or exhibiting low-quality corporate cultures. Less readable disclosures can impact stakeholder decision-making and undermine confidence in the related disclosures of firms. This creates socio-economic implications, as ineffective communication may hinder the societal transition to sustainability or mislead investors and regulators regarding the true impact of climate change on firms’ operations. Furthermore, the study underscores the importance of implementing measures to mitigate exposure to climate change.

We acknowledge the potential limitations of the current study. First, the unobserved heterogeneity among firms may impact our results. We mitigate this possibility by controlling for a wide range of observable firm-level variables. Additionally, we recognize that, due to data availability, our study focuses on a U.S. sample and relies on the bog index. Lastly, despite our efforts to address alternative interpretations of the results, we cannot completely eliminate the possibility that the complexity associated with climate change disclosure may contribute to the reduced readability of annual reports.

This study opens up various avenues for future research. For instance, subsequent studies could investigate the market implications of decreased readability resulting from exposure to climate change. Furthermore, this study focuses on one aspect of managerial decisions regarding financial reporting: the readability of annual reports. Future research could assess the impact of climate change on decisions related to other attributes of financial reporting, utilizing data from various contexts beyond the U.S. Lastly, future studies could analyze the readability of Environmental, Social, and Governance (ESG) reports to provide deeper insights into the influence of climate change on other forms of disclosure.