The Causal Effect of Intangible Cultural Heritage on Corporate Performance—Evidence from a Double Machine Learning Model

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript investigates the causal effect of intangible cultural heritage (ICH) on firm performance, addressing a notable gap in the literature. The topic demonstrates originality and academic relevance, particularly by focusing on how informal cultural institutions, such as ICH, may influence corporate outcomes. Overall, the study presents an interesting contribution; however, several aspects of the manuscript require clarification, restructuring, and methodological enhancement to strengthen the validity and impact of the findings. My detailed comments are as follows:

1.The abstract should present representative quantitative results to support the findings. The authors are encouraged to incorporate key numerical evidence in the abstract to illustrate the core conclusions, such as the causal effect of ICH on ROA.

2.The research objective is not clearly stated in the abstract. It is recommended that the authors explicitly articulate the purpose of the study around Line 12, ensuring the description is anchored to the relationship between ROA and ICH.

3.The background section (Lines 27–47) requires closer alignment with the study’s focus. Attention should be paid to the writing hierarchy, accurately describing the relevant research context and gaps. The background should revolve around the core theme in cultural economics of “how informal institutions influence firm behavior,” highlighting the uniqueness of ICH before providing further context. A reorganization of the logical flow is suggested.

4.The logical coherence of the literature review needs improvement. Section 2.1 discusses “the influence of culture on corporate risk-taking,” while Section 2.2 abruptly shifts to “the governance effects of ICH” without any transitional sentence. It is recommended that the discussion logically flow from generalized cultural effects to the specific governance effects of ICH.

5.The research gap (Lines 77–83) should not be embedded within a subsection of the literature review. Instead, it should appear as a standalone summary paragraph, outlining the existing deficiencies in research on ICH.

6.It is suggested to extend the current double machine learning (DML) model by incorporating a spatial weight matrix to construct a spatial DML model and conduct robustness checks. This would address potential spatial autocorrelation of firm performance induced by regional ICH variables and enhance the rigor of causal inference.
6.In Section 3.2.3 (Control Variables), the definition of “Con” (Lines 160–161) is inconsistent with the variable name, causing conceptual confusion.

7.The logic of Section 6.2, Limitations and Avenues for Future Research, is unclear. It is recommended that the limitations and future research directions be separated and their internal structure clarified. Moreover, the practical implications should not be discussed in a generic way; targeted strategies based on the specific empirical results should be proposed to highlight the study’s applied value.

8.The discussion in Sections 5.1–5.3 (Lines 352–407) is weakly connected to the empirical findings. The current discussion does not provide in-depth theoretical explanations for key results and largely reiterates the conclusions. The authors are advised to reorganize the discussion around the main findings, offer deeper theoretical interpretations, and explicitly compare their results with those of existing studies to better demonstrate the study’s contribution.

9.The paper introduces a comparison between ICH and the ideologically oriented Long March Spirit; however, the subsequent discussion merely states that “the Long March Spirit has no significant effect” without exploring its potential value in non-financial domains. Additionally, the background section (Line 48) does not provide sufficient justification for introducing this comparison. The authors should clearly explain why the Long March Spirit is included as a comparative framework and ensure that the discussion section substantively addresses this point to maintain conceptual consistency.

Author Response

Manuscript ID: sustainability-4140929

Title: The Causal Effect of Intangible Cultural Heritage on Corporate Performance—Evidence from a Double Machine Learning Model

Authors: Fusheng Xie, Yuxia Wang 

 

 

Dear Reviewer,

 

Thank you for your instructions to revise the manuscript. Please attached herewith find our revised manuscript of the above referred paper for your consideration and publication in Sustainability. We have thoroughly revised and modified the manuscript according to the comments of the reviewers. Detailed responses to the comments are listed below point by point.

 

We appreciate the valuable comments. Line number are provided in responses to the comments (if needed) and highlighted text where changes are made in the paper text.

 

   Response to Reviewer:

1. The abstract should present representative quantitative results to support the findings. The authors are encouraged to incorporate key numerical evidence in the abstract to illustrate the core conclusions, such as the causal effect of ICH on ROA.

Leveraging the Double Machine Learning (DML) method, we find that ICH exerts a significant positive effect on ROA, with the estimated coefficient ranging from 0.0018 to 0.0020 (p < 0.01). In contrast, the effect of the Spirit of the Long March is statistically insignificant.  (Line 16-Line 19)

2. The research objective is not clearly stated in the abstract. It is recommended that the authors explicitly articulate the purpose of the study around Line 12, ensuring the description is anchored to the relationship between ROA and ICH.

This study focuses on place-based social culture, specifically using a firm's local Intangible Cultural Heritage (ICH) as the core cultural variable. We aim to identify its causal effect on firm performance (measured by ROA) and contrast it with another pivotal cultural force—the Spirit of the Long March.  (Line 12-Line 16)

3. The background section (Lines 27–47) requires closer alignment with the study’s focus. Attention should be paid to the writing hierarchy, accurately describing the relevant research context and gaps. The background should revolve around the core theme in cultural economics of “how informal institutions influence firm behavior,” highlighting the uniqueness of ICH before providing further context. A reorganization of the logical flow is suggested.

A foundational tenet of institutional economics posits that informal institutions—particularly culture—exert a profound influence on economic development and organizational efficacy by shaping individual mental models and behavioral norms (North, 1990) [1]. Within the domain of corporate finance, this proposition has spawned a rich spectrum of inquiry. Early research focused on national-level cultural dimensions (Hofstede, 2001) [2], revealing the broad impact of macro-cultural values on corporate governance. In recent years, scholarly perspective has shifted towards greater granularity, examining how specific, operationalizable cultural forms—such as religious traditions (Hilary & Hui, 2009) [3] and geographically-bounded social capital (Li et al., 2013) [4]—concretely guide strategic choices and risk preferences of firms. However, a fundamental challenge persists in this field: establishing robust causal linkages between specific cultural elements and corporate financial performance after controlling for complex confounders, and systematically unveiling the underlying channels of influence.

To address this challenge with precision, this study focuses on a specific, place-based dimension of social culture: the local Intangible Cultural Heritage (ICH). We define ICH as the living traditions, practices, and knowledge embedded within a firm’s geographic locale, distinct from internal organizational culture. Our objective is to examine how this external, locally-rooted informal institution causally influences firm performance.

China presents a valuable social laboratory for deepening this inquiry. Its unique institutional environment blends a long-standing traditional civilizational heritage with a modern revolutionary political culture. Existing research has progressed along two distinct veins: one exploring the governance functions of traditional philosophies, exemplified by Confucianism, on corporate behaviors such as curbing over-investment and fostering innovation (Chen et al., 2019; Yan et al., 2021) [5] [6]; the other examining the shaping influence of revolutionary culture (e.g., "red" traditions) on corporate financial policies like cash holdings and leverage (Li et al., 2023; Zhang & Zheng, 2025) [7] [8]. Nevertheless, a crucial yet long-overlooked cultural form is Intangible Cultural Heritage (ICH). ICH does not constitute an abstract ideological system but rather encompasses "living" practical knowledge embedded in local crafts, rituals, and oral traditions, transmitted intergenerationally through master-apprentice relationships and community interactions. It represents a mode of "practical governance" rooted in everyday productive life, whose potential economic and corporate finance value remains largely unexplored in the extant literature.

This oversight highlights four interrelated research gaps: First, there is a lack of theoretical attention and empirical examination regarding how embodied, skill-oriented informal institutionslike ICH influence the operations of modern firms. Second, cultural studies are generally plagued by endogeneity issues, such as omitted variable bias and reverse causality, casting doubt on the reliability of causal inferences. Third, the specific transmission mechanisms through which culture affects performance often remain a "black box," with few studies clearly validating the financial pathways. Fourth, existing literature frequently treats culture as a homogeneous concept, lacking comparative analysis of the governance functions of different cultural types, thereby obscuring the distinct operational logics of various cultural forms.

This study aims to address these shortcomings systematically. We focus on ICH to investigate its causal effect on firm performance (measured by Return on Assets, ROA). To more clearly delineate the governance attributes of ICH, we introduce the Spirit of the Long March as a comparative referent. The Spirit of the Long March represents an ideological revolutionary culture formed in modern China, emphasizing perseverance, discipline, and collectivism. Through this contrast, we construct a theoretical comparative framework to test whether the economic efficacy of culture varies according to its inherent nature and transmission logic ("practical transmission" vs. "value internalization"). To overcome the core difficulties in causal identification, we employ the advanced Double Machine Learning (DML) estimation method (Chernozhukov et al., 2018) [9]. DML can flexibly handle high-dimensional control variables and complex non-linear relationships, thereby more effectively isolating confounding factors and providing a cleaner, more robust estimate of the cultural effects of interest.

The contributions of this study are threefold. Methodologically, we pioneer the application of the DML framework in cultural finance, offering a more rigorous and powerful analytical tool to address the field's inherent endogeneity challenges, thereby enhancing the credibility of causal inference. Theoretically, we introduce ICH as a key independent variable into corporate finance research for the first time and innovatively propose a dichotomous cultural typology based on "utilitarian-practical" versus "ideological-revolutionary" attributes, deepening theoretical understanding of how heterogeneity in informal institutions shapes corporate behavior. Mechanistically, we move beyond the generalized discussions of prior research by empirically testing and confirming that financial conservatism (specifically manifested as reduced corporate leverage) serves as a key mediating pathway through which ICH enhances firm performance, thereby partially illuminating the financial "black box" of cultural influence on corporate decision-making.  (Line 31-Line 102)

4. The logical coherence of the literature review needs improvement. Section 2.1 discusses “the influence of culture on corporate risk-taking,” while Section 2.2 abruptly shifts to “the governance effects of ICH” without any transitional sentence. It is recommended that the discussion logically flow from generalized cultural effects to the specific governance effects of ICH.
5. The research objective is not clearly stated in the abstract. It is recommended that the authors explicitly articulate the purpose of the study around Line 12, ensuring the description is anchored to the relationship between ROA and ICH.

2.1 Informal Institutions, Culture, and Corporate Governance

Institutional economic theory posits that informal constraints (e.g., customs, traditions, culture), together with formal rules, constitute an institutional matrix that shapes long-term economic performance (North, 1990) [1]. In the micro-realm of corporate finance, the mechanisms through which culture—as a deep-seated informal institution—exerts its influence have been increasingly scrutinized. Early research primarily employed cross-country data to examine the effects of macro-level indicators, such as Hofstede's cultural dimensions, on corporate governance and risk-taking (Hofstede, 2001) [2]. To mitigate confounding factors arising from country-level heterogeneity, recent literature has shifted focus towards subnational cultures or specific cultural elements within countries, such as geographically-based religious adherence (Hilary & Hui, 2009) [3] or regional social capital (Li et al., 2013) [4], aiming to identify a cleaner effect of culture on corporate decisions. Collectively, these studies reveal that culture primarily influences strategic and financial choices by shaping managers' cognitive frameworks, risk preferences, and ethical norms ( Guiso et al. , 2006) [10].

China, with its profound traditional civilization and unique modern historical trajectory, provides a fertile context for exploring cultural governance. Existing research has converged along two distinct strands: one focuses on traditional philosophical culture, exemplified by Confucianism, demonstrating its role in curbing corporate over-investment (Chen et al., 2019, Ma et al., 2025) [5] [11] and stimulating innovation inputs (Yan et al., 2021;Huang et al., 2024) [6][12]; the other examines ideological culture rooted in revolutionary history, finding its influence on corporate cash-holding policies (Li et al., 2023) [7] and financial leverage decisions (Zhang & Zheng, 2025) [8]. However, a parallel cultural system—Intangible Cultural Heritage (ICH), which is grounded in localized, practice-based knowledge—and its potential corporate governance and financial implications remain a blind spot in mainstream literature.

Synthesizing this body of work reveals four critical and interconnected gaps that collectively underpin the motivation for this study. First, there exists a significant missing variable problem: ICH, as a "living" and practice-oriented informal institution, has not yet been systematically integrated into the analytical framework of corporate finance. Second, research in this domain grapples with a persistent endogeneity conundrum, as cultural variables are frequently correlated with unobserved regional or firm-level characteristics, thereby casting doubt on the reliability of causal inferences. Third, the black box of mechanisms remains largely unopened—the specific financial channels through which culture influences firm performance are still predominantly theoretical and lack rigorous econometric validation. Finally, the literature suffers from a deficiency in typology, often treating “culture” as a homogeneous construct rather than distinguishing and comparing the governance logics and economic effects of different cultural types, such as utility- or skill-based traditions versus ideology- or value-driven forms.

2.2 The Governance Effects of Intangible Cultural Heritage

Although direct studies are scarce, integrating institutional theory, the resource-based view, and social capital theory allows for the construction of a three-tiered framework to understand the governance effects of ICH.

(1) Cognitive and Behavioral Shaping Mechanism

The core carrier of ICH is the "artisan spirit," emphasizing the pursuit of technical excellence, persistent dedication, and reverence for quality (Xiao & Liu, 2015) [13]. Through community ethos and socialization, this spiritual core subtly shapes the mental models of managers operating within these regions, cultivating a decision-making preference characterized by long-term orientation and prudence. Behavioral finance indicates that managers' personal traits and local cultural environments profoundly influence their risk perception (Malmendier & Nagel, 2011) [14]. The principles of "continuous improvement" and "century-long inheritance" embedded in ICH are likely to reduce managers' tolerance for high-volatility, high-leverage financial strategies, prompting more conservative financing and investment policies.

(2) Resource and Signaling Mechanism

From the resource-based view perspective, ICH can be regarded as a unique strategic resource—local cultural capital—possessing value, rarity, and inimitability. Firms located in ICH-rich areas can embed these distinctive cultural symbols into product design, brand narratives, and corporate identity (Ginesti et al., 2018) [15]. This not only facilitates product differentiation and premium pricing but also builds profound customer loyalty and brand moats based on cultural identification. This unique competitive advantage can directly enhance the sustainability of profitability and potentially reduce firms' reliance on high-risk, short-term debt financing.

(3) Informal Monitoring and Constraint Mechanism

The survival of ICH depends on tight-knit community networks and shared industry norms, which naturally incubate high levels of social capital and reputation-based constraint mechanisms (Dong et al., 2018) [16]. Firms operating within such communities face informal monitoring from peers, inheritor groups, and the local public. Any myopic or risky behavior (e.g., aggressive financial manipulation or excessive leverage) that could jeopardize the firm's long-term survival and, by extension, tarnish the local cultural reputation, will incur higher social sanction costs. This invisible community pressure constitutes an effective informal governance mechanism, constraining managerial opportunism and guiding the adoption of socially responsible and more financially sound policies.

Preliminary empirical evidence supports the economic influence of ICH. For instance, Pan et al. (2023) [17] found that artistic ICH significantly promotes corporate innovation, revealing its role in enhancing long-term dynamic capabilities. Based on the above theoretical reasoning and initial evidence, we propose the first core hypothesis:

H1: Intangible Cultural Heritage (ICH) has a positive impact on corporate performance.

2.3 The Spirit of the Long March and Corporate Performance

In contrast to the "utilitarian" nature of ICH, the Spirit of the Long March represents a significant form of ideological revolutionary culture in modern China, emphasizing arduous struggle, discipline, and collectivism. Existing literature presents a theoretical tension regarding its impact on corporate financial performance. on one hand, its inherent discipline may strengthen internal control and improve operational efficiency (Wang et al., 2023; Wu et al.,2024) [18] [19]; on the other hand, its advocacy of sacrifice and collective priority may lead firms to deviate somewhat from pure profit maximization, shifting focus towards political and social responsibilities (Shen et al., 2025) [20].

We argue that this tension stems from the fundamental difference in the inherent attributes and operational logics between the Spirit of the Long March and ICH. The Spirit of the Long March is a political ethic inculcated through propaganda and education, centered on shaping beliefs and loyalty; whereas ICH is a practical knowledge system transmitted intergenerationally through skill imitation and community practice, centered on enhancing skills and sustaining livelihoods. This divergence suggests their primary channels of influence differ. The governance effect of the Spirit of the Long March should be more pronounced in non-financial domains, such as strengthening corporate compliance (Wu & Hong, 2024) [19], enhancing ESG performance (Shen et al., 2025) [20], or increasing organizational resilience, rather than directly affecting accounting profitability (ROA) under normal circumstances. Any potential impact on ROA is more likely to be indirect, context-dependent, and statistically difficult to detect. Therefore, we propose the following comparative hypothesis:

H2: The Spirit of the Long March has no significant direct impact on corporate performance (ROA).

2.4 The Mediating Role of Financial Conservatism

To illuminate the "black box" of cultural influence on performance, we further investigate specific transmission channels. Based on the "artisan spirit" and "sustainable development" philosophy embedded in ICH (Dai, 2025) [21] , we posit that financial conservatism (specifically manifested in prudent capital structure decisions) is a key mediating mechanism through which ICH enhances performance.

The long-term orientation and risk-aversion culture inherent in ICH becomes internalized as a firm's pursuit of financial soundness. The financial risks, agency costs, and constraints on investment flexibility associated with excessive debt (Myers, 1977) [22] contradict the sustainable and self-controlled philosophy advocated by ICH. Consequently, firms more deeply immersed in ICH culture are expected to exhibit a lower preference for financial leverage. Extensive empirical research has robustly confirmed that high leverage is generally detrimental to firm performance (Frank & Goyal, 2009; Akhtaret al., 2022; Yang al.,2016) [23] [24] [25]. Therefore, ICH may indirectly improve firm performance by guiding firms to adopt more conservative financing policies (i.e., reducing leverage), thereby lowering financial distress costs and interest burdens. This leads to our mechanism hypothesis:

H3: Intangible Cultural Heritage enhances corporate performance indirectly by reducing corporate leverage.    

6. It is suggested to extend the current double machine learning (DML) model by incorporating a spatial weight matrix to construct a spatial DML model and conduct robustness checks. This would address potential spatial autocorrelation of firm performance induced by regional ICH variables and enhance the rigor of causal inference.

Due to constraints in the available data, which lack detailed geospatial information, a formal spatial analysis could not be conducted in the current study. We have incorporated relevant robustness checks to address this limitation. Future iterations of this work will integrate a Spatial DML model to formally explore spatial dimensions.

4.6 Robustness test

To address potential endogeneity concerns and to further validate the robustness of our core findings, we conduct a series of sensitivity analyses. A primary critique in cultural economics is that observed correlations between cultural variables and firm outcomes may be driven by unobserved heterogeneity or model misspecification. Specifically, 1) omitted variable bias may arise if factors influencing both regional cultural endowment and corporate performance are not adequately controlled for; 2) spatial or temporal confounding might inflate the estimated cultural effect; and 3) the choice of machine learning estimators and control variable sets could influence the stability of the results. In response to these methodological concerns—and as suggested during the review process—we systematically examine the sensitivity of our results to alternative model specifications, including the incorporation of additional control variables, year-fixed effects, and diverse machine learning algorithms. These tests are designed not only to confirm the reliability of our causal claims but also to demonstrate that the positive effect of Intangible Cultural Heritage (ICH) is not an artifact of particular modeling choices or neglected confounders.

To further enhance the robustness of our causal inference and account for potential confounding from macroeconomic shocks and time-specific trends that may simultaneously affect regional cultural vibrancy and corporate performance, we incorporate year-fixed effects into our DML model. The control set now expands to W={Lev,Size,Age,Ins,Con,Ind,Soe,I.year}, where I.year denotes a set of year dummies.

Table 11 presents the estimation results for both cultural variables under different machine learning estimators after including year-fixed effects. The coefficient for Culture (ICH) remains positive and statistically significant at the 1% level across all four methods, with estimates ranging from 0.0019 to 0.0020. This stability, in both magnitude and significance, reinforces the reliability of our core finding. Notably, the point estimate shows a slight increase relative to the baseline specifications without time effects (e.g., from 0.0018 to approximately 0.0020). This suggests that after purging the influence of aggregate time-varying factors (e.g., business cycles, nationwide policy shifts), the genuine positive effect of ICH on firm performance is, if anything, more pronounced.

Conversely, the coefficient for Treat (the Spirit of the Long March) remains statistically insignificant across all specifications, with point estimates close to zero. This result further corroborates Hypothesis H2, indicating the absence of a direct financial performance effect from this ideological cultural force, even after controlling for temporal shocks.

Table 11  Robustness Test: Controlling for Time-fixed Effects

	Method (rf)	Method (lassocv)	Method (gradboost)	Method (nnet)
Treat	-0.0003	-0.0006	-0.0006	-0.0007
	(0.0011)	(0.0011)	(0.0010)	(0.0010)
Culture	0.0019***	0.0020***	0.0019***	0.0020***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)
N	23052	23052	23052	23052

 

The robustness tests presented in Table 12 yield several consistent and noteworthy conclusions regarding the causal effects of cultural variables on firm performance. First and foremost, the positive impact of Intangible Cultural Heritage (ICH) on Return on Assets (ROA) demonstrates a high degree of robustness across multiple model specifications. The coefficient for Culture remains statistically significant at the 1% or 5% levels under all four machine learning estimators—Random Forest, LASSO, Gradient Boosting, and Neural Networks—and across four distinct sets of control variables. Although the point estimates exhibit a slight decline, from approximately 0.0017 to around 0.0009, as additional controls such as audit quality (Big4), growth opportunities (TobinQ), and ownership structure (Soe) are incrementally included, the significance and positive direction of the effect remain stable. This pattern suggests that while factors like growth prospects and firm ownership may partially mediate or confound the relationship, the independent, positive influence of ICH on firm performance is robustly preserved. These results provide strong and consistent support for Hypothesis H1.

In clear contrast, the coefficient for the Spirit of the Long March (Treat) is statistically insignificant across nearly all model configurations. Its estimates are consistently close to zero and lack statistical significance, with only isolated instances of marginal significance at the 10% level in certain specifications. This systematic absence of a significant direct effect strongly corroborates Hypothesis H2, indicating that this form of ideological revolutionary culture does not exert a measurable direct impact on corporate financial performance as captured by ROA.

Furthermore, the stability of these findings across diverse machine learning methods underscores that the conclusions are not sensitive to the choice of estimation technique. The consistent results produced by models with differing underlying assumptions and functional forms—from tree-based ensembles to regularized linear models and neural networks—enhance the credibility of the causal inferences. The gradual attenuation of the ICH coefficient with the inclusion of TobinQ and Soe points to the nuanced nature of the relationship, hinting that part of ICH's beneficial effect may operate through channels such as enhancing growth opportunities or being more effectively leveraged within certain ownership structures, a fruitful avenue for further mechanistic exploration.

In summary, the battery of robustness checks confirms that the promotive effect of ICH on firm performance is a reliable finding, while the null effect of the Long March Spirit is equally stable. The core results withstand variations in both the set of control variables and the machine learning methodology employed for estimation.

Table 12 Robustness of Main Results Across Different Machine Learning Methods and Control Variable Sets

Model	Treat (Long March Spirit)	Culture (Intangible Heritage)	Covariates
Random Forest	-0.0009 (0.0011)	0.0017***(0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind
LASSO	-0.0009 (0.0011)	0.0017***(0.0004)
Gradient Boost	-0.0017 (0.0011)	0.0018***(0.0004)
Neural Network	-0.0017 (0.0010)	0.0019*** (0.0004)
Random Forest	-0.0006 (0.0011)	0.0017***(0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4
LASSO	-0.0008 (0.0011)	0.0017***(0.0004)
Gradient Boost	-0.0014 (0.0011)	0.0018***(0.0004)
Neural Network	-0.0015 (0.0011)	0.0018***(0.0004)
Random Forest	-0.0014 (0.0010)	0.0009** (0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4,Tobinq
LASSO	-0.0016(0.0010)	0.0009** (0.0004)
Gradient Boost	-0.0018* (0.0010)	0.0010***(0.0004)
Neural Network	-0.0019* (0.0010)	0.0010** (0.0004)
Random Forest	-0.0014 (0.0011)	0.0009** (0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4,Tobinq,Soe
LASSO	-0.0014 (0.0011)	0.0008** (0.0004)
Gradient Boost	-0.0017 (0.0011)	0.0010***(0.0004)
Neural Network	-0.0018* (0.0011)	0.0010*** (0.0004)

Note: Standard errors in parentheses; * p<0.10, ** p<0.05, *** p<0.01.

   

7. In Section 3.2.3 (Control Variables), the definition of “Con” (Lines 160–161) is inconsistent with the variable name, causing conceptual confusion.

Thank you for your careful review. We have revised the definition of the control variable "Con" in lines 160–161 to align it precisely with the variable name, eliminating any previous inconsistency and enhancing the clarity of our variable description.

8. The logic of Section 6.2, Limitations and Avenues for Future Research, is unclear. It is recommended that the limitations and future research directions be separated and their internal structure clarified. Moreover, the practical implications should not be discussed in a generic way; targeted strategies based on the specific empirical results should be proposed to highlight the study’s applied value.

6.2. Limitations and Avenues for Future Research

This study acknowledges several limitations, which also illuminate promising directions for future inquiry.

6.2.1 Limitations

Macro-level Variable Measurement. This study employs the number of Intangible Cultural Heritage (ICH) projects at the city level as a proxy for the regional cultural environment. While this approach effectively measures the concentration of cultural resources across regions, it cannot precisely capture the extent to which individual firms actively utilize, internalize, or strategically prioritize these ICH resources.

Challenges in Causal Identification. Although the Double Machine Learning method effectively controls for high-dimensional observable confounders, unobservable omitted variables—such as difficult-to-quantify local informal institutions, unique social capital, or tacit knowledge networks—may still pose a potential threat to the causal estimates.

Neglect of Spatial Dependence. This study assumes spatial independence in firm performance across the sample. However, in reality, factors like cultural diffusion, regional economic linkages, or policy imitation may lead to spatial autocorrelation in firm performance. Ignoring this spatial dependency structure could result in misestimated standard errors, thereby compromising the rigor of statistical inference.

Incomplete Mechanism Exploration. This study confirms financial conservatism (deleveraging) as one effective mediating pathway, but this is likely only one among multiple parallel mechanisms. The potential impacts of ICH on corporate brand value, stakeholder trust, and even craftsmanship-driven innovation have not been fully examined, leaving the complete picture of the transmission mechanisms to be fully mapped.

6.2.2 Avenues for Future Research

Building upon these limitations, future research can advance in the following directions:

Developing Micro-level Measurement Tools. Future studies could attempt to construct firm-level indicators of cultural embeddedness or utilization by employing textual analysis (mining references to specific cultural elements in annual reports or CSR reports), survey questionnaires (measuring the cultural values of management and employees), or case studies. This would deepen the measurement from "regional cultural environment" to "firm-specific cultural capital."

Expanding Causal Identification Strategies. To mitigate omitted variable bias, future research could seek cleaner exogenous policy shocks for identification, such as exploiting the quasi-natural experiment of a successful designation on the "Representative List of the Intangible Cultural Heritage of Humanity." Concurrently, future work should develop and apply a "Spatial Double Machine Learning" model. By incorporating a spatial weight matrix, this model would explicitly control for the spatial spillover effects of both performance and culture, thereby yielding more robust causal estimates while accommodating complex functional forms and high-dimensional controls.

Systematically Testing Multiple Mechanisms. It is recommended to adopt more systematic mediation analysis frameworks (e.g., multiple mediation models or causal mediation analysis) to concurrently test the various parallel pathways through which ICH might affect firm performance—such as "brand enhancement," "trust building," and "innovation incentive"—and to quantify the relative importance of each path.

9. The discussion in Sections 5.1–3 (Lines 352–407) is weakly connected to the empirical findings. The current discussion does not provide in-depth theoretical explanations for key results and largely reiterates the conclusions. The authors are advised to reorganize the discussion around the main findings, offer deeper theoretical interpretations, and explicitly compare their results with those of existing studies to better demonstrate the study’s contribution.

5.1 Theoretical Interpretation and Contributions of Key Findings

This study leverages a Double Machine Learning model to provide robust evidence for the causal effect of Intangible Cultural Heritage (ICH) on corporate financial performance. The core findings indicate that ICH has a significant positive impact on firm ROA, while the Spirit of the Long March shows no significant direct effect. These results not only validate our main hypotheses but also offer new theoretical insights for interdisciplinary research at the intersection of cultural economics and corporate finance.

First, the positive effect of ICH deepens our understanding of "informal institutions as corporate governance mechanisms." Unlike the universal values emphasized in traditional literature, such as religion or Confucianism (Guiso et al., 2006; Hilary & Hui, 2009) [3] [10], ICH represents a unique form of localized, practical, and embedded cultural capital. Our results suggest that this cultural form, transmitted through community practices, apprenticeships, and practical wisdom, can effectively shape managers' risk perceptions and long-term orientation (Xiao & Liu, 2015) [13], thereby translating into more prudent corporate decisions. This aligns with the conclusions of Li et al. (2013) [4] regarding the influence of culture on risk-taking but advances the analysis from national cultural dimensions to more specific and actionable local cultural heritage.

Second, the divergence in effects between different types of culture (utilitarian traditional culture vs. ideological revolutionary culture) introduces an important theoretical distinction. Our finding that the Spirit of the Long March has no significant impact on ROA differs from some studies focusing on the effects of revolutionary culture on specific financial behaviors (e.g., reducing cash holdings) (Li et al., 2023; Zhang & Zheng, 2025) [7] [8]. This discrepancy may stem from the theoretical logic of the "domain specificity" of cultural influence. As a political ideology, the governance function of the Long March Spirit may be more focused on shaping normative behaviors (e.g., discipline, collectivism) and non-financial objectives (e.g., ESG performance, crisis resilience) (Shen et al., 2025) [20] rather than directly driving short-term accounting profits. This aligns with North's (1990) [1] discussion on the scope of institutional constraints on behavior. Therefore, through comparative analysis, this study advocates that future cultural research should move beyond the assumption that "culture is a monolithic construct" and instead explore the differentiated governance pathways and boundary conditions of different cultural types.

5.2 Exploring the Mechanisms and Boundary Conditions of Heterogeneous Findings

The heterogeneity analysis reveals the boundary conditions under which the ICH effect operates, providing important contextualization for theory.

The stronger ICH effect in State-Owned Enterprises (SOEs) can be explained by theories of institutional isomorphism and resource complementarity. Due to their ownership nature, SOEs are more deeply embedded in local socio-political networks and have a natural affinity with ICH as a form of local social capital. This institutional fit allows SOEs to more effectively absorb and translate the trust networks and reputation mechanisms inherent in ICH (Dong et al., 2018) [16], leading to more pronounced performance improvements. This finding complements the existing literature on how ownership structure moderates cultural effects (Chen et al., 2019) [5], suggesting that in the presence of strongly embedded informal institutions like ICH, SOEs' "disadvantages" may transform into unique resource advantages.

The amplifying effect of older CEOs on the ICH effect resonates with the Upper Echelons Theory perspective on how managers' cognitive bases influence strategic choices. Older CEOs typically possess richer local social experience, deeper internalization of traditional values, and a stronger concern for long-term reputation. This makes them more adept at interpreting and utilizing the "artisan spirit" and sustainable development philosophy conveyed by ICH and translating them into conservative financial policies (reducing leverage). This finding extends cultural influence research from the macro-regional level to the micro-level of managerial characteristics, revealing a "human capital channel" for cultural value conversion.

5.3 Theoretical Significance of the Financial Conservatism Mediation Mechanism

The mechanism analysis confirms that reducing financial leverage is a key transmission channel through which ICH enhances performance. This finding has dual theoretical significance.

On one hand, it opens the "black box" of how culture influences corporate financial decision-making. Previous studies often stopped at correlational conclusions about "culture affecting performance," lacking tests of the intermediate economic decision-making processes. We demonstrate that the values of focus, prudence, and risk aversion advocated by ICH are concretely manifested in capital structure choices (Myers, 1977) [22], namely lower debt ratios. This provides a solid empirical link in the "culture-behavior-performance" chain.

On the other hand, it expands capital structure theory. Beyond traditional explanations based on trade-off theory, pecking order theory, and market perfection, our research shows that local informal cultural norms are an important yet overlooked factor influencing corporate leverage decisions. Firms do not make financial decisions in a vacuum; the "cultural air" they inhabit subtly shapes management's risk preferences and financing inclinations.

 

10. The paper introduces a comparison between ICH and the ideologically oriented Long March Spirit; however, the subsequent discussion merely states that “the Long March Spirit has no significant effect” without exploring its potential value in non-financial domains. Additionally, the background section (Line 48) does not provide sufficient justification for introducing this comparison. The authors should clearly explain why the Long March Spirit is included as a comparative framework and ensure that the discussion section substantively addresses this point to maintain conceptual consistency.

This study aims to address these shortcomings systematically. We focus on ICH to investigate its causal effect on firm performance (measured by Return on Assets, ROA). To more clearly delineate the governance attributes of ICH, we introduce the Spirit of the Long March as a comparative referent. The Spirit of the Long March represents an ideological revolutionary culture formed in modern China, emphasizing perseverance, discipline, and collectivism. Through this contrast, we construct a theoretical comparative framework to test whether the economic efficacy of culture varies according to its inherent nature and transmission logic ("practical transmission" vs. "value internalization"). To overcome the core difficulties in causal identification, we employ the advanced Double Machine Learning (DML) estimation method (Chernozhukov et al., 2018) [9]. DML can flexibly handle high-dimensional control variables and complex non-linear relationships, thereby more effectively isolating confounding factors and providing a cleaner, more robust estimate of the cultural effects of interest.   (Line 76 -Line 88)

2.3 The Spirit of the Long March and Corporate Performance

In contrast to the "utilitarian" nature of ICH, the Spirit of the Long March represents a significant form of ideological revolutionary culture in modern China, emphasizing arduous struggle, discipline, and collectivism. Existing literature presents a theoretical tension regarding its impact on corporate financial performance. on one hand, its inherent discipline may strengthen internal control and improve operational efficiency (Wang et al., 2023; Wu et al.,2024) [18] [19]; on the other hand, its advocacy of sacrifice and collective priority may lead firms to deviate somewhat from pure profit maximization, shifting focus towards political and social responsibilities (Shen et al., 2025) [20].

We argue that this tension stems from the fundamental difference in the inherent attributes and operational logics between the Spirit of the Long March and ICH. The Spirit of the Long March is a political ethic inculcated through propaganda and education, centered on shaping beliefs and loyalty; whereas ICH is a practical knowledge system transmitted intergenerationally through skill imitation and community practice, centered on enhancing skills and sustaining livelihoods. This divergence suggests their primary channels of influence differ. The governance effect of the Spirit of the Long March should be more pronounced in non-financial domains, such as strengthening corporate compliance (Wu & Hong, 2024) [19], enhancing ESG performance (Shen et al., 2025) [20], or increasing organizational resilience, rather than directly affecting accounting profitability (ROA) under normal circumstances. Any potential impact on ROA is more likely to be indirect, context-dependent, and statistically difficult to detect. Therefore, we propose the following comparative hypothesis:

H2: The Spirit of the Long March has no significant direct impact on corporate performance (ROA).

 

 

Please contact me if you have any questions on the revised manuscript.

 

Best regards,

 

Fusheng Xie

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The article entitled The Causal Effect of Intangible Cultural Heritage on Corporate Performance - Evidence from a Double Machine Learning Model is well written and contributes to the research area on the impact of ICH on corporate financial performance. The entire study is well explained and consistent in its research process and results. The authors explained in detail how they obtained the outcome. Below are some comments on each part that are recommended for improvement, not mistakes.

Abstract: Since the beginning (the abstract) of the article, it is not so clear what culture the authors mean – the organisational culture or the social culture of a specific country or location, and the authors recommend that they be more precise.

Introduction, Literature review, and hypothesis part. The introduction section reviews the literature on how norms, beliefs, and religion influence corporate decision-making. This part also addresses the research gap based on the literature review. In my opinion, the additional literature review section is unnecessary, especially since this review is not based on a systematic literature review. My suggestion for authors is to combine the first and second sections into an introduction and to end it with a clear statement of the research goal and hypothesis. In this part, there is also one of the main omissions. The authors address the research gap but do not explain why they consider the Spirit of the Long March as a treatment. It seems that some kind of mental shortcut has been used here, perhaps obvious in China, but probably not recognisable to readers from other countries. As the article is intended for an international audience, I recommend explaining why the Spirit of the Long March is a significant factor in the geographic variation of Intangible Culture.

The Research Design section is very well prepared. All the data is explained. The only suggestion to authors is to ensure that no other geographical differences could influence the analysis results, such as differences in local policy, funding, and regional power, which can affect intangible cultural heritage. Using three causal discovery algorithms is a good practice for identifying the set of controlled variables.

The empirical findings: The analysis was well explained and based on the approach described in the literature. The authors provide a sound justification for the selection of subsequent steps in the study and the decisions they made before moving on to the next stage of analysis. This makes it clear how the subsequent results were obtained. Given such justification for limitations and decisions, the results obtained can be considered valid and support the hypothesis. The discussion section is also well prepared and references other research. The manuscript's language is clear and understandable.

Author Response

Manuscript ID: sustainability-4140929

Title: The Causal Effect of Intangible Cultural Heritage on Corporate Performance—Evidence from a Double Machine Learning Model

Authors: Fusheng Xie, Yuxia Wang

 

 

Dear Reviewer,

 

Thank you for your instructions to revise the manuscript. Please attached herewith find our revised manuscript of the above referred paper for your consideration and publication in Sustainability. We have thoroughly revised and modified the manuscript according to the comments of the reviewers. Detailed responses to the comments are listed below point by point.

 

We appreciate the valuable comments. Line number are provided in responses to the comments (if needed) and highlighted text where changes are made in the paper text.

 

   Response to Reviewer:

1. Abstract: Since the beginning (the abstract) of the article, it is not so clear what culture the authors mean – the organisational culture or the social culture of a specific country or location, and the authors recommend that they be more precise.

While culture is recognized as a foundational informal institution, quantifying its causal impact on corporate outcomes remains a challenge. This study focuses on place-based social culture, specifically using a firm's local Intangible Cultural Heritage (ICH) as the core cultural variable. We aim to identify its causal effect on firm performance (measured by ROA) and contrast it with another pivotal cultural force—the Spirit of the Long March. Leveraging the Double Machine Learning (DML) method, we find that ICH exerts a significant positive effect on ROA, with the estimated coefficient ranging from 0.0018 to 0.0020 (p < 0.01). In contrast, the effect of the Spirit of the Long March is statistically insignificant. Cross-validation using multiple machine learning algorithms confirms the robustness of our findings. Heterogeneity analysis reveals that the effect of ICH is more pronounced in state-owned enterprises and firms with older CEOs. Mechanism analysis further uncovers that ICH enhances performance primarily by fostering financial conservatism, specifically through a reduction in corporate leverage. Our study provides robust causal evidence on how place-based traditional culture shapes corporate financial policies and underscores the economic value of preserving intangible cultural assets.   (Line 11- Line 25)

To address this challenge with precision, this study focuses on a specific, place-based dimension of social culture: the local Intangible Cultural Heritage (ICH). We define ICH as the living traditions, practices, and knowledge embedded within a firm’s geographic locale, distinct from internal organizational culture. Our objective is to examine how this external, locally-rooted informal institution causally influences firm performance.   (Line 45- Line 49)

 

2. Introduction, Literature review, and hypothesis part. The introduction section reviews the literature on how norms, beliefs, and religion influence corporate decision-making. This part also addresses the research gap based on the literature review. In my opinion, the additional literature review section is unnecessary, especially since this review is not based on a systematic literature review. My suggestion for authors is to combine the first and second sections into an introduction and to end it with a clear statement of the research goal and hypothesis. In this part, there is also one of the main omissions. The authors address the research gap but do not explain why they consider the Spirit of the Long March as a treatment. It seems that some kind of mental shortcut has been used here, perhaps obvious in China, but probably not recognisable to readers from other countries. As the article is intended for an international audience, I recommend explaining why the Spirit of the Long March is a significant factor in the geographic variation of Intangible Culture.
3. Introduction

A foundational tenet of institutional economics posits that informal institutions—particularly culture—exert a profound influence on economic development and organizational efficacy by shaping individual mental models and behavioral norms (North, 1990) [1]. Within the domain of corporate finance, this proposition has spawned a rich spectrum of inquiry. Early research focused on national-level cultural dimensions (Hofstede, 2001) [2], revealing the broad impact of macro-cultural values on corporate governance. In recent years, scholarly perspective has shifted towards greater granularity, examining how specific, operationalizable cultural forms—such as religious traditions (Hilary & Hui, 2009) [3] and geographically-bounded social capital (Li et al., 2013) [4]—concretely guide strategic choices and risk preferences of firms. However, a fundamental challenge persists in this field: establishing robust causal linkages between specific cultural elements and corporate financial performance after controlling for complex confounders, and systematically unveiling the underlying channels of influence.

To address this challenge with precision, this study focuses on a specific, place-based dimension of social culture: the local Intangible Cultural Heritage (ICH). We define ICH as the living traditions, practices, and knowledge embedded within a firm’s geographic locale, distinct from internal organizational culture. Our objective is to examine how this external, locally-rooted informal institution causally influences firm performance.

China presents a valuable social laboratory for deepening this inquiry. Its unique institutional environment blends a long-standing traditional civilizational heritage with a modern revolutionary political culture. Existing research has progressed along two distinct veins: one exploring the governance functions of traditional philosophies, exemplified by Confucianism, on corporate behaviors such as curbing over-investment and fostering innovation (Chen et al., 2019; Yan et al., 2021) [5] [6]; the other examining the shaping influence of revolutionary culture (e.g., "red" traditions) on corporate financial policies like cash holdings and leverage (Li et al., 2023; Zhang & Zheng, 2025) [7] [8]. Nevertheless, a crucial yet long-overlooked cultural form is Intangible Cultural Heritage (ICH). ICH does not constitute an abstract ideological system but rather encompasses "living" practical knowledge embedded in local crafts, rituals, and oral traditions, transmitted intergenerationally through master-apprentice relationships and community interactions. It represents a mode of "practical governance" rooted in everyday productive life, whose potential economic and corporate finance value remains largely unexplored in the extant literature.

This oversight highlights four interrelated research gaps: First, there is a lack of theoretical attention and empirical examination regarding how embodied, skill-oriented informal institutionslike ICH influence the operations of modern firms. Second, cultural studies are generally plagued by endogeneity issues, such as omitted variable bias and reverse causality, casting doubt on the reliability of causal inferences. Third, the specific transmission mechanisms through which culture affects performance often remain a "black box," with few studies clearly validating the financial pathways. Fourth, existing literature frequently treats culture as a homogeneous concept, lacking comparative analysis of the governance functions of different cultural types, thereby obscuring the distinct operational logics of various cultural forms.

This study aims to address these shortcomings systematically. We focus on ICH to investigate its causal effect on firm performance (measured by Return on Assets, ROA). To more clearly delineate the governance attributes of ICH, we introduce the Spirit of the Long March as a comparative referent. The Spirit of the Long March represents an ideological revolutionary culture formed in modern China, emphasizing perseverance, discipline, and collectivism. Through this contrast, we construct a theoretical comparative framework to test whether the economic efficacy of culture varies according to its inherent nature and transmission logic ("practical transmission" vs. "value internalization"). To overcome the core difficulties in causal identification, we employ the advanced Double Machine Learning (DML) estimation method (Chernozhukov et al., 2018) [9]. DML can flexibly handle high-dimensional control variables and complex non-linear relationships, thereby more effectively isolating confounding factors and providing a cleaner, more robust estimate of the cultural effects of interest.

The contributions of this study are threefold. Methodologically, we pioneer the application of the DML framework in cultural finance, offering a more rigorous and powerful analytical tool to address the field's inherent endogeneity challenges, thereby enhancing the credibility of causal inference. Theoretically, we introduce ICH as a key independent variable into corporate finance research for the first time and innovatively propose a dichotomous cultural typology based on "utilitarian-practical" versus "ideological-revolutionary" attributes, deepening theoretical understanding of how heterogeneity in informal institutions shapes corporate behavior. Mechanistically, we move beyond the generalized discussions of prior research by empirically testing and confirming that financial conservatism (specifically manifested as reduced corporate leverage) serves as a key mediating pathway through which ICH enhances firm performance, thereby partially illuminating the financial "black box" of cultural influence on corporate decision-making.

The remainder of this paper is structured as follows: Section II reviews the relevant literature and develops the research hypotheses. Section III elaborates on the research design, data sources, and DML model specification. Section IV reports the empirical results, including tests of the main effects, robustness checks, heterogeneity analysis,  mechanism validation test and robustness test. Section V provides an in-depth discussion of the findings, explaining their theoretical implications and practical significance. Section VI concludes the paper, outlines its limitations, and suggests directions for future research.

2. Literature Review and Research Hypotheses

2.1 Informal Institutions, Culture, and Corporate Governance

Institutional economic theory posits that informal constraints (e.g., customs, traditions, culture), together with formal rules, constitute an institutional matrix that shapes long-term economic performance (North, 1990) [1]. In the micro-realm of corporate finance, the mechanisms through which culture—as a deep-seated informal institution—exerts its influence have been increasingly scrutinized. Early research primarily employed cross-country data to examine the effects of macro-level indicators, such as Hofstede's cultural dimensions, on corporate governance and risk-taking (Hofstede, 2001) [2]. To mitigate confounding factors arising from country-level heterogeneity, recent literature has shifted focus towards subnational cultures or specific cultural elements within countries, such as geographically-based religious adherence (Hilary & Hui, 2009) [3] or regional social capital (Li et al., 2013) [4], aiming to identify a cleaner effect of culture on corporate decisions. Collectively, these studies reveal that culture primarily influences strategic and financial choices by shaping managers' cognitive frameworks, risk preferences, and ethical norms ( Guiso et al. , 2006) [10].

China, with its profound traditional civilization and unique modern historical trajectory, provides a fertile context for exploring cultural governance. Existing research has converged along two distinct strands: one focuses on traditional philosophical culture, exemplified by Confucianism, demonstrating its role in curbing corporate over-investment (Chen et al., 2019, Ma et al., 2025) [5] [11] and stimulating innovation inputs (Yan et al., 2021;Huang et al., 2024) [6][12]; the other examines ideological culture rooted in revolutionary history, finding its influence on corporate cash-holding policies (Li et al., 2023) [7] and financial leverage decisions (Zhang & Zheng, 2025) [8]. However, a parallel cultural system—Intangible Cultural Heritage (ICH), which is grounded in localized, practice-based knowledge—and its potential corporate governance and financial implications remain a blind spot in mainstream literature.

Synthesizing this body of work reveals four critical and interconnected gaps that collectively underpin the motivation for this study. First, there exists a significant missing variable problem: ICH, as a "living" and practice-oriented informal institution, has not yet been systematically integrated into the analytical framework of corporate finance. Second, research in this domain grapples with a persistent endogeneity conundrum, as cultural variables are frequently correlated with unobserved regional or firm-level characteristics, thereby casting doubt on the reliability of causal inferences. Third, the black box of mechanisms remains largely unopened—the specific financial channels through which culture influences firm performance are still predominantly theoretical and lack rigorous econometric validation. Finally, the literature suffers from a deficiency in typology, often treating “culture” as a homogeneous construct rather than distinguishing and comparing the governance logics and economic effects of different cultural types, such as utility- or skill-based traditions versus ideology- or value-driven forms.

2.2 The Governance Effects of Intangible Cultural Heritage

Although direct studies are scarce, integrating institutional theory, the resource-based view, and social capital theory allows for the construction of a three-tiered framework to understand the governance effects of ICH.

(1) Cognitive and Behavioral Shaping Mechanism

The core carrier of ICH is the "artisan spirit," emphasizing the pursuit of technical excellence, persistent dedication, and reverence for quality (Xiao & Liu, 2015) [13]. Through community ethos and socialization, this spiritual core subtly shapes the mental models of managers operating within these regions, cultivating a decision-making preference characterized by long-term orientation and prudence. Behavioral finance indicates that managers' personal traits and local cultural environments profoundly influence their risk perception (Malmendier & Nagel, 2011) [14]. The principles of "continuous improvement" and "century-long inheritance" embedded in ICH are likely to reduce managers' tolerance for high-volatility, high-leverage financial strategies, prompting more conservative financing and investment policies.

(2) Resource and Signaling Mechanism

From the resource-based view perspective, ICH can be regarded as a unique strategic resource—local cultural capital—possessing value, rarity, and inimitability. Firms located in ICH-rich areas can embed these distinctive cultural symbols into product design, brand narratives, and corporate identity (Ginesti et al., 2018) [15]. This not only facilitates product differentiation and premium pricing but also builds profound customer loyalty and brand moats based on cultural identification. This unique competitive advantage can directly enhance the sustainability of profitability and potentially reduce firms' reliance on high-risk, short-term debt financing.

(3) Informal Monitoring and Constraint Mechanism

The survival of ICH depends on tight-knit community networks and shared industry norms, which naturally incubate high levels of social capital and reputation-based constraint mechanisms (Dong et al., 2018) [16]. Firms operating within such communities face informal monitoring from peers, inheritor groups, and the local public. Any myopic or risky behavior (e.g., aggressive financial manipulation or excessive leverage) that could jeopardize the firm's long-term survival and, by extension, tarnish the local cultural reputation, will incur higher social sanction costs. This invisible community pressure constitutes an effective informal governance mechanism, constraining managerial opportunism and guiding the adoption of socially responsible and more financially sound policies.

Preliminary empirical evidence supports the economic influence of ICH. For instance, Pan et al. (2023) [17] found that artistic ICH significantly promotes corporate innovation, revealing its role in enhancing long-term dynamic capabilities. Based on the above theoretical reasoning and initial evidence, we propose the first core hypothesis:

H1: Intangible Cultural Heritage (ICH) has a positive impact on corporate performance.

2.3 The Spirit of the Long March and Corporate Performance

In contrast to the "utilitarian" nature of ICH, the Spirit of the Long March represents a significant form of ideological revolutionary culture in modern China, emphasizing arduous struggle, discipline, and collectivism. Existing literature presents a theoretical tension regarding its impact on corporate financial performance. on one hand, its inherent discipline may strengthen internal control and improve operational efficiency (Wang et al., 2023; Wu et al.,2024) [18] [19]; on the other hand, its advocacy of sacrifice and collective priority may lead firms to deviate somewhat from pure profit maximization, shifting focus towards political and social responsibilities (Shen et al., 2025) [20].

We argue that this tension stems from the fundamental difference in the inherent attributes and operational logics between the Spirit of the Long March and ICH. The Spirit of the Long March is a political ethic inculcated through propaganda and education, centered on shaping beliefs and loyalty; whereas ICH is a practical knowledge system transmitted intergenerationally through skill imitation and community practice, centered on enhancing skills and sustaining livelihoods. This divergence suggests their primary channels of influence differ. The governance effect of the Spirit of the Long March should be more pronounced in non-financial domains, such as strengthening corporate compliance (Wu & Hong, 2024) [19], enhancing ESG performance (Shen et al., 2025) [20], or increasing organizational resilience, rather than directly affecting accounting profitability (ROA) under normal circumstances. Any potential impact on ROA is more likely to be indirect, context-dependent, and statistically difficult to detect. Therefore, we propose the following comparative hypothesis:

H2: The Spirit of the Long March has no significant direct impact on corporate performance (ROA).

2.4 The Mediating Role of Financial Conservatism

To illuminate the "black box" of cultural influence on performance, we further investigate specific transmission channels. Based on the "artisan spirit" and "sustainable development" philosophy embedded in ICH (Dai, 2025) [21] , we posit that financial conservatism (specifically manifested in prudent capital structure decisions) is a key mediating mechanism through which ICH enhances performance.

The long-term orientation and risk-aversion culture inherent in ICH becomes internalized as a firm's pursuit of financial soundness. The financial risks, agency costs, and constraints on investment flexibility associated with excessive debt (Myers, 1977) [22] contradict the sustainable and self-controlled philosophy advocated by ICH. Consequently, firms more deeply immersed in ICH culture are expected to exhibit a lower preference for financial leverage. Extensive empirical research has robustly confirmed that high leverage is generally detrimental to firm performance (Frank & Goyal, 2009; Akhtaret al., 2022; Yang al.,2016) [23] [24] [25]. Therefore, ICH may indirectly improve firm performance by guiding firms to adopt more conservative financing policies (i.e., reducing leverage), thereby lowering financial distress costs and interest burdens. This leads to our mechanism hypothesis:

H3: Intangible Cultural Heritage enhances corporate performance indirectly by reducing corporate leverage.

3. The Research Design section is very well prepared. All the data is explained. The only suggestion to authors is to ensure that no other geographical differences could influence the analysis results, such as differences in local policy, funding, and regional power, which can affect intangible cultural heritage. Using three causal discovery algorithms is a good practice for identifying the set of controlled variables.

We thank you for this constructive suggestion. Following your advice, we have incorporated time-fixed effects and relevant control variables into the robustness checks. Due to methodological constraints related to data availability, we were unable to control for all the specific variables mentioned in your comment within the current framework. We acknowledge this limitation and will prioritize the collection of the requisite data to address these factors in our future research.

4.6 Robustness test

To address potential endogeneity concerns and to further validate the robustness of our core findings, we conduct a series of sensitivity analyses. A primary critique in cultural economics is that observed correlations between cultural variables and firm outcomes may be driven by unobserved heterogeneity or model misspecification. Specifically, 1) omitted variable bias may arise if factors influencing both regional cultural endowment and corporate performance are not adequately controlled for; 2) spatial or temporal confounding might inflate the estimated cultural effect; and 3) the choice of machine learning estimators and control variable sets could influence the stability of the results. In response to these methodological concerns—and as suggested during the review process—we systematically examine the sensitivity of our results to alternative model specifications, including the incorporation of additional control variables, year-fixed effects, and diverse machine learning algorithms. These tests are designed not only to confirm the reliability of our causal claims but also to demonstrate that the positive effect of Intangible Cultural Heritage (ICH) is not an artifact of particular modeling choices or neglected confounders.

To further enhance the robustness of our causal inference and account for potential confounding from macroeconomic shocks and time-specific trends that may simultaneously affect regional cultural vibrancy and corporate performance, we incorporate year-fixed effects into our DML model. The control set now expands to W={Lev,Size,Age,Ins,Con,Ind,Soe,I.year}, where I.year denotes a set of year dummies.

Table 11 presents the estimation results for both cultural variables under different machine learning estimators after including year-fixed effects. The coefficient for Culture (ICH) remains positive and statistically significant at the 1% level across all four methods, with estimates ranging from 0.0019 to 0.0020. This stability, in both magnitude and significance, reinforces the reliability of our core finding. Notably, the point estimate shows a slight increase relative to the baseline specifications without time effects (e.g., from 0.0018 to approximately 0.0020). This suggests that after purging the influence of aggregate time-varying factors (e.g., business cycles, nationwide policy shifts), the genuine positive effect of ICH on firm performance is, if anything, more pronounced.

Conversely, the coefficient for Treat (the Spirit of the Long March) remains statistically insignificant across all specifications, with point estimates close to zero. This result further corroborates Hypothesis H2, indicating the absence of a direct financial performance effect from this ideological cultural force, even after controlling for temporal shocks.

Table 11  Robustness Test: Controlling for Time-fixed Effects

	Method (rf)	Method (lassocv)	Method (gradboost)	Method (nnet)
Treat	-0.0003	-0.0006	-0.0006	-0.0007
	(0.0011)	(0.0011)	(0.0010)	(0.0010)
Culture	0.0019***	0.0020***	0.0019***	0.0020***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)
N	23052	23052	23052	23052

Note: Standard errors in parentheses; * p<0.10, ** p<0.05, *** p<0.01.

The robustness tests presented in Table 12 yield several consistent and noteworthy conclusions regarding the causal effects of cultural variables on firm performance. First and foremost, the positive impact of Intangible Cultural Heritage (ICH) on Return on Assets (ROA) demonstrates a high degree of robustness across multiple model specifications. The coefficient for Culture remains statistically significant at the 1% or 5% levels under all four machine learning estimators—Random Forest, LASSO, Gradient Boosting, and Neural Networks—and across four distinct sets of control variables. Although the point estimates exhibit a slight decline, from approximately 0.0017 to around 0.0009, as additional controls such as audit quality (Big4), growth opportunities (TobinQ), and ownership structure (Soe) are incrementally included, the significance and positive direction of the effect remain stable. This pattern suggests that while factors like growth prospects and firm ownership may partially mediate or confound the relationship, the independent, positive influence of ICH on firm performance is robustly preserved. These results provide strong and consistent support for Hypothesis H1.

In clear contrast, the coefficient for the Spirit of the Long March (Treat) is statistically insignificant across nearly all model configurations. Its estimates are consistently close to zero and lack statistical significance, with only isolated instances of marginal significance at the 10% level in certain specifications. This systematic absence of a significant direct effect strongly corroborates Hypothesis H2, indicating that this form of ideological revolutionary culture does not exert a measurable direct impact on corporate financial performance as captured by ROA.

Furthermore, the stability of these findings across diverse machine learning methods underscores that the conclusions are not sensitive to the choice of estimation technique. The consistent results produced by models with differing underlying assumptions and functional forms—from tree-based ensembles to regularized linear models and neural networks—enhance the credibility of the causal inferences. The gradual attenuation of the ICH coefficient with the inclusion of TobinQ and Soe points to the nuanced nature of the relationship, hinting that part of ICH's beneficial effect may operate through channels such as enhancing growth opportunities or being more effectively leveraged within certain ownership structures, a fruitful avenue for further mechanistic exploration.

In summary, the battery of robustness checks confirms that the promotive effect of ICH on firm performance is a reliable finding, while the null effect of the Long March Spirit is equally stable. The core results withstand variations in both the set of control variables and the machine learning methodology employed for estimation.

Table 12 Robustness of Main Results Across Different Machine Learning Methods and Control Variable Sets

Model	Treat (Long March Spirit)	Culture (Intangible Heritage)	Covariates
Random Forest	-0.0009 (0.0011)	0.0017***(0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind
LASSO	-0.0009 (0.0011)	0.0017***(0.0004)
Gradient Boost	-0.0017 (0.0011)	0.0018***(0.0004)
Neural Network	-0.0017 (0.0010)	0.0019*** (0.0004)
Random Forest	-0.0006 (0.0011)	0.0017***(0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4
LASSO	-0.0008 (0.0011)	0.0017***(0.0004)
Gradient Boost	-0.0014 (0.0011)	0.0018***(0.0004)
Neural Network	-0.0015 (0.0011)	0.0018***(0.0004)
Random Forest	-0.0014 (0.0010)	0.0009** (0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4,Tobinq
LASSO	-0.0016(0.0010)	0.0009** (0.0004)
Gradient Boost	-0.0018* (0.0010)	0.0010***(0.0004)
Neural Network	-0.0019* (0.0010)	0.0010** (0.0004)
Random Forest	-0.0014 (0.0011)	0.0009** (0.0004)	Basic Set is Lev, Size, Age, Ins, Con, Ind, Big4,Tobinq,Soe
LASSO	-0.0014 (0.0011)	0.0008** (0.0004)
Gradient Boost	-0.0017 (0.0011)	0.0010***(0.0004)
Neural Network	-0.0018* (0.0011)	0.0010*** (0.0004)

Note: Standard errors in parentheses; * p<0.10, ** p<0.05, *** p<0.01.

H3: Intangible Cultural Heritage enhances corporate performance indirectly by reducing corporate leverage.

 

Please contact me if you have any questions on the revised manuscript.

 

Best regards,

 

Fusheng Xie

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study employs a Double Machine Learning (DML) framework to identify the causal effect of Intangible Cultural Heritage (ICH) on the performance of Chinese A-share listed firms, as measured by return on assets (ROA), and conducts a comparative analysis with revolutionary culture represented by the Spirit of the Long March. The topic is novel and the data and methodology are generally solid; however, there remains substantial room for improvement in the theoretical conceptualization of cultural variables, the clarification of the scope and limits of causal identification, the rigor of the mechanism analysis, and the depth of engagement with the existing literature.

1.Introduction

The review of the existing literature lacks sufficient depth, and the articulation of the research gap remains inadequately substantiated. The statement of contributions is not sufficiently explicit. It is recommended that the authors conduct a more thorough synthesis of the cultural finance literature and, in light of the practice-oriented nature of ICH, more rigorously justify the theoretical and practical significance of examining the impact of ICH on corporate performance. The manuscript should explicitly clarify its marginal contributions relative to the existing “culture–performance” literature, especially in terms of enhanced causal identification and the testability of underlying mechanisms.

2. Literature Review and Research Hypotheses

2.3 H2 The conceptual definition and differentiation between the Spirit of the Long March and ICH are overly simplistic and insufficiently supported by relevant theories or prior literature. It is suggested that the authors engage more deeply with institutional theory and related cultural or ideological frameworks to strengthen the theoretical dialogue.

3. Research Design

3.3 The details of the model specification are insufficiently reported, and the stability as well as the potential risk of misidentification associated with the causal discovery algorithms are not adequately discussed. Key implementation details and tuning parameters for both the DML framework and the causal discovery algorithms are not specified. It is recommended that the authors provide essential information on parameter settings and implementation procedures. In addition, robustness checks for the causal discovery results should be reported.

4. Empirical Findings

4.3 The interpretation of the main results lacks sufficient depth. The results primarily report statistical significance while failing to provide a systematic discussion of their economic implications. It is recommended that discussions on economic significance be incorporated, along with comparisons against the effect sizes documented in existing studies on cultural finance.

4.5 The mechanism analysis is not sufficiently comprehensive. The current analysis only validates the mediating role of leverage without formally testing the statistical significance of the mediation effect, nor does it consider alternative plausible channels through which ICH may affect corporate performance. It is recommended that the authors clearly specify the methodology used to compute the mediation effect and its associated statistical inference framework. Moreover, the discussion should emphasize the “partial explanatory” nature of the proposed mechanism to avoid overgeneralization of the mediating pathway.

5. Discussion

The discussion of policy implications remains overly macro-level and is insufficiently linked to actionable recommendations at the firm level. It is advisable to translate the policy implications into more concrete and operational guidance for corporate practice.

6. Conclusion

Although limitations are acknowledged, the reflection on the core proxy variables and the underlying causal identification assumptions is not sufficiently in-depth. It is recommended that the limitations section be further elaborated and more clearly connected to future research directions.

Author Response

Manuscript ID: sustainability-4140929

Title: The Causal Effect of Intangible Cultural Heritage on Corporate Performance—Evidence from a Double Machine Learning Model

Authors: Fusheng Xie, Yuxia Wang

 

 

Dear Reviewer,

 

Thank you for your instructions to revise the manuscript. Please attached herewith find our revised manuscript of the above referred paper for your consideration and publication in Sustainability. We have thoroughly revised and modified the manuscript according to the comments of the reviewers. Detailed responses to the comments are listed below point by point.

 

We appreciate the valuable comments. Line number are provided in responses to the comments (if needed) and highlighted text where changes are made in the paper text.

 

   Response to Reviewer:

1. The review of the existing literature lacks sufficient depth, and the articulation of the research gap remains inadequately substantiated. The statement of contributions is not sufficiently explicit. It is recommended that the authors conduct a more thorough synthesis of the cultural finance literature and, in light of the practice-oriented nature of ICH, more rigorously justify the theoretical and practical significance of examining the impact of ICH on corporate performance. The manuscript should explicitly clarify its marginal contributions relative to the existing “culture–performance” literature, especially in terms of enhanced causal identification and the testability of underlying mechanisms.
2. Introduction

A foundational tenet of institutional economics posits that informal institutions—particularly culture—exert a profound influence on economic development and organizational efficacy by shaping individual mental models and behavioral norms (North, 1990) [1]. Within the domain of corporate finance, this proposition has spawned a rich spectrum of inquiry. Early research focused on national-level cultural dimensions (Hofstede, 2001) [2], revealing the broad impact of macro-cultural values on corporate governance. In recent years, scholarly perspective has shifted towards greater granularity, examining how specific, operationalizable cultural forms—such as religious traditions (Hilary & Hui, 2009) [3] and geographically-bounded social capital (Li et al., 2013) [4]—concretely guide strategic choices and risk preferences of firms. However, a fundamental challenge persists in this field: establishing robust causal linkages between specific cultural elements and corporate financial performance after controlling for complex confounders, and systematically unveiling the underlying channels of influence.

To address this challenge with precision, this study focuses on a specific, place-based dimension of social culture: the local Intangible Cultural Heritage (ICH). We define ICH as the living traditions, practices, and knowledge embedded within a firm’s geographic locale, distinct from internal organizational culture. Our objective is to examine how this external, locally-rooted informal institution causally influences firm performance.

China presents a valuable social laboratory for deepening this inquiry. Its unique institutional environment blends a long-standing traditional civilizational heritage with a modern revolutionary political culture. Existing research has progressed along two distinct veins: one exploring the governance functions of traditional philosophies, exemplified by Confucianism, on corporate behaviors such as curbing over-investment and fostering innovation (Chen et al., 2019; Yan et al., 2021) [5] [6]; the other examining the shaping influence of revolutionary culture (e.g., "red" traditions) on corporate financial policies like cash holdings and leverage (Li et al., 2023; Zhang & Zheng, 2025) [7] [8]. Nevertheless, a crucial yet long-overlooked cultural form is Intangible Cultural Heritage (ICH). ICH does not constitute an abstract ideological system but rather encompasses "living" practical knowledge embedded in local crafts, rituals, and oral traditions, transmitted intergenerationally through master-apprentice relationships and community interactions. It represents a mode of "practical governance" rooted in everyday productive life, whose potential economic and corporate finance value remains largely unexplored in the extant literature.

This oversight highlights four interrelated research gaps: First, there is a lack of theoretical attention and empirical examination regarding how embodied, skill-oriented informal institutionslike ICH influence the operations of modern firms. Second, cultural studies are generally plagued by endogeneity issues, such as omitted variable bias and reverse causality, casting doubt on the reliability of causal inferences. Third, the specific transmission mechanisms through which culture affects performance often remain a "black box," with few studies clearly validating the financial pathways. Fourth, existing literature frequently treats culture as a homogeneous concept, lacking comparative analysis of the governance functions of different cultural types, thereby obscuring the distinct operational logics of various cultural forms.

This study aims to address these shortcomings systematically. We focus on ICH to investigate its causal effect on firm performance (measured by Return on Assets, ROA). To more clearly delineate the governance attributes of ICH, we introduce the Spirit of the Long March as a comparative referent. The Spirit of the Long March represents an ideological revolutionary culture formed in modern China, emphasizing perseverance, discipline, and collectivism. Through this contrast, we construct a theoretical comparative framework to test whether the economic efficacy of culture varies according to its inherent nature and transmission logic ("practical transmission" vs. "value internalization"). To overcome the core difficulties in causal identification, we employ the advanced Double Machine Learning (DML) estimation method (Chernozhukov et al., 2018) [9]. DML can flexibly handle high-dimensional control variables and complex non-linear relationships, thereby more effectively isolating confounding factors and providing a cleaner, more robust estimate of the cultural effects of interest.

The contributions of this study are threefold. Methodologically, we pioneer the application of the DML framework in cultural finance, offering a more rigorous and powerful analytical tool to address the field's inherent endogeneity challenges, thereby enhancing the credibility of causal inference. Theoretically, we introduce ICH as a key independent variable into corporate finance research for the first time and innovatively propose a dichotomous cultural typology based on "utilitarian-practical" versus "ideological-revolutionary" attributes, deepening theoretical understanding of how heterogeneity in informal institutions shapes corporate behavior. Mechanistically, we move beyond the generalized discussions of prior research by empirically testing and confirming that financial conservatism (specifically manifested as reduced corporate leverage) serves as a key mediating pathway through which ICH enhances firm performance, thereby partially illuminating the financial "black box" of cultural influence on corporate decision-making.

The remainder of this paper is structured as follows: Section II reviews the relevant literature and develops the research hypotheses. Section III elaborates on the research design, data sources, and DML model specification. Section IV reports the empirical results, including tests of the main effects, robustness checks, heterogeneity analysis,  mechanism validation test and robustness test. Section V provides an in-depth discussion of the findings, explaining their theoretical implications and practical significance. Section VI concludes the paper, outlines its limitations, and suggests directions for future research.

2.1 Informal Institutions, Culture, and Corporate Governance

Institutional economic theory posits that informal constraints (e.g., customs, traditions, culture), together with formal rules, constitute an institutional matrix that shapes long-term economic performance (North, 1990) [1]. In the micro-realm of corporate finance, the mechanisms through which culture—as a deep-seated informal institution—exerts its influence have been increasingly scrutinized. Early research primarily employed cross-country data to examine the effects of macro-level indicators, such as Hofstede's cultural dimensions, on corporate governance and risk-taking (Hofstede, 2001) [2]. To mitigate confounding factors arising from country-level heterogeneity, recent literature has shifted focus towards subnational cultures or specific cultural elements within countries, such as geographically-based religious adherence (Hilary & Hui, 2009) [3] or regional social capital (Li et al., 2013) [4], aiming to identify a cleaner effect of culture on corporate decisions. Collectively, these studies reveal that culture primarily influences strategic and financial choices by shaping managers' cognitive frameworks, risk preferences, and ethical norms ( Guiso et al. , 2006) [10].

China, with its profound traditional civilization and unique modern historical trajectory, provides a fertile context for exploring cultural governance. Existing research has converged along two distinct strands: one focuses on traditional philosophical culture, exemplified by Confucianism, demonstrating its role in curbing corporate over-investment (Chen et al., 2019, Ma et al., 2025) [5] [11] and stimulating innovation inputs (Yan et al., 2021;Huang et al., 2024) [6][12]; the other examines ideological culture rooted in revolutionary history, finding its influence on corporate cash-holding policies (Li et al., 2023) [7] and financial leverage decisions (Zhang & Zheng, 2025) [8]. However, a parallel cultural system—Intangible Cultural Heritage (ICH), which is grounded in localized, practice-based knowledge—and its potential corporate governance and financial implications remain a blind spot in mainstream literature.

Synthesizing this body of work reveals four critical and interconnected gaps that collectively underpin the motivation for this study. First, there exists a significant missing variable problem: ICH, as a "living" and practice-oriented informal institution, has not yet been systematically integrated into the analytical framework of corporate finance. Second, research in this domain grapples with a persistent endogeneity conundrum, as cultural variables are frequently correlated with unobserved regional or firm-level characteristics, thereby casting doubt on the reliability of causal inferences. Third, the black box of mechanisms remains largely unopened—the specific financial channels through which culture influences firm performance are still predominantly theoretical and lack rigorous econometric validation. Finally, the literature suffers from a deficiency in typology, often treating “culture” as a homogeneous construct rather than distinguishing and comparing the governance logics and economic effects of different cultural types, such as utility- or skill-based traditions versus ideology- or value-driven forms.

2. 3 H2 The conceptual definition and differentiation between the Spirit of the Long March and ICH are overly simplistic and insufficiently supported by relevant theories or prior literature. It is suggested that the authors engage more deeply with institutional theory and related cultural or ideological frameworks to strengthen the theoretical dialogue.

2.3 The Spirit of the Long March and Corporate Performance

In contrast to the "utilitarian" nature of ICH, the Spirit of the Long March represents a significant form of ideological revolutionary culture in modern China, emphasizing arduous struggle, discipline, and collectivism. Existing literature presents a theoretical tension regarding its impact on corporate financial performance. on one hand, its inherent discipline may strengthen internal control and improve operational efficiency (Wang et al., 2023; Wu et al.,2024) [18] [19]; on the other hand, its advocacy of sacrifice and collective priority may lead firms to deviate somewhat from pure profit maximization, shifting focus towards political and social responsibilities (Shen et al., 2025) [20].

We argue that this tension stems from the fundamental difference in the inherent attributes and operational logics between the Spirit of the Long March and ICH. The Spirit of the Long March is a political ethic inculcated through propaganda and education, centered on shaping beliefs and loyalty; whereas ICH is a practical knowledge system transmitted intergenerationally through skill imitation and community practice, centered on enhancing skills and sustaining livelihoods. This divergence suggests their primary channels of influence differ. The governance effect of the Spirit of the Long March should be more pronounced in non-financial domains, such as strengthening corporate compliance (Wu & Hong, 2024) [19], enhancing ESG performance (Shen et al., 2025) [20], or increasing organizational resilience, rather than directly affecting accounting profitability (ROA) under normal circumstances. Any potential impact on ROA is more likely to be indirect, context-dependent, and statistically difficult to detect. Therefore, we propose the following comparative hypothesis:

H2: The Spirit of the Long March has no significant direct impact on corporate performance (ROA).   (LIne 193-Line 218)

3. 3 The details of the model specification are insufficiently reported, and the stability as well as the potential risk of misidentification associated with the causal discovery algorithms are not adequately discussed. Key implementation details and tuning parameters for both the DML framework and the causal discovery algorithms are not specified. It is recommended that the authors provide essential information on parameter settings and implementation procedures. In addition, robustness checks for the causal discovery results should be reported.

 3.3.4 implement environment

To ensure transparent and robust identification of confounders, we employed three established causal discovery algorithms: the constraint-based PC algorithm, the FCI algorithm (which accounts for potential latent confounders), and the score-based GES algorithm. These algorithms were implemented using the causal-learn library in Python. For the subsequent causal estimation, we applied a Double Machine Learning (DML) framework through the ddml package in Stata, adopting a 5-fold cross-fitting design to mitigate overfitting and enhance the robustness of the estimator. Throughout both the causal discovery and DML estimation stages, all relevant algorithmic parameters were maintained at their default system values to ensure reproducibility and minimize subjective tuning.   (Line 333-Line 343)

4. 3 The interpretation of the main results lacks sufficient depth. The results primarily report statistical significance while failing to provide a systematic discussion of their economic implications. It is recommended that discussions on economic significance be incorporated, along with comparisons against the effect sizes documented in existing studies on cultural finance.

While the statistical significance of ICH is clear, assessing its economic magnitude is crucial. Our DML estimate indicates that a one-unit increase in the ICH index is associated with an increase in ROA of about 0.16 to 0.20 percentage points (coefficient range: 0.0016–0.0020). To contextualize this effect, consider that the standard deviation of ROA in our sample is 0.064. The ICH effect thus accounts for roughly 2.8% to 3.3% of one standard deviation change in ROA. This magnitude is economically meaningful. For comparison, seminal studies in cultural finance find effects of similar order (Hilary & Hui (2009) [3]. Our estimated effect for ICH is also comparable to the documented impact of certain internal governance reforms in emerging markets. This confirms that ICH is not merely a statistical artifact but a cultural force with tangible, non-trivial consequences for firm profitability. In contrast, the consistently null effect for the Treat variable (Long March Spirit) aligns with our theoretical expectation and underscores that not all cultural forms directly translate into short-term financial performance.   (Line 403-Line 415)

5. 5 The mechanism analysis is not sufficiently comprehensive. The current analysis only validates the mediating role of leverage without formally testing the statistical significance of the mediation effect, nor does it consider alternative plausible channels through which ICH may affect corporate performance. It is recommended that the authors clearly specify the methodology used to compute the mediation effect and its associated statistical inference framework. Moreover, the discussion should emphasize the “partial explanatory” nature of the proposed mechanism to avoid overgeneralization of the mediating pathway.

4.5 Mechanism Analysis

Building on H3, we posit that financial conservatism (proxied by lower leverage, Lev) is a key channel through which ICH enhances ROA. We go beyond simple correlation by formally testing the statistical significance of the mediation effect using a DML-based mediation framework. This method extends the DML logic to estimate the Average Causal Mediation Effect—the portion of the total effect that operates through the mediator—while robustly controlling for confounders.  (Line 453- Line 458)

This finding provides robust, statistically-validated support for H3. However, we emphasize the partial explanatory nature of this mechanism. The proportion mediated is roughly 66% suggests that while reducing leverage is an important pathway, other plausible channels are likely at play. These may include, but are not limited to: enhanced product quality and brand value stemming from the "artisan spirit" (resource-signaling channel), improved stakeholder trust and employee morale fostered by community-based social capital (informal governance channel), or increased innovation as shown by Pan et al. (2023) [17]. Future research should formally model and test these alternative channels. Our analysis therefore establishes financial conservatism as one validated, significant mechanism, but cautions against overgeneralizing it as the exclusive pathway linking ICH to corporate performance.     (line 474-Line 485)

6. The discussion of policy implications remains overly macro-level and is insufficiently linked to actionable recommendations at the firm level. It is advisable to translate the policy implications into more concrete and operational guidance for corporate practice.
7. Discussion

5.1 Theoretical Interpretation and Contributions of Key Findings

This study leverages a Double Machine Learning model to provide robust evidence for the causal effect of Intangible Cultural Heritage (ICH) on corporate financial performance. The core findings indicate that ICH has a significant positive impact on firm ROA, while the Spirit of the Long March shows no significant direct effect. These results not only validate our main hypotheses but also offer new theoretical insights for interdisciplinary research at the intersection of cultural economics and corporate finance.

First, the positive effect of ICH deepens our understanding of "informal institutions as corporate governance mechanisms." Unlike the universal values emphasized in traditional literature, such as religion or Confucianism (Guiso et al., 2006; Hilary & Hui, 2009) [3] [10], ICH represents a unique form of localized, practical, and embedded cultural capital. Our results suggest that this cultural form, transmitted through community practices, apprenticeships, and practical wisdom, can effectively shape managers' risk perceptions and long-term orientation (Xiao & Liu, 2015) [13], thereby translating into more prudent corporate decisions. This aligns with the conclusions of Li et al. (2013) [4] regarding the influence of culture on risk-taking but advances the analysis from national cultural dimensions to more specific and actionable local cultural heritage.

Second, the divergence in effects between different types of culture (utilitarian traditional culture vs. ideological revolutionary culture) introduces an important theoretical distinction. Our finding that the Spirit of the Long March has no significant impact on ROA differs from some studies focusing on the effects of revolutionary culture on specific financial behaviors (e.g., reducing cash holdings) (Li et al., 2023; Zhang & Zheng, 2025) [7] [8]. This discrepancy may stem from the theoretical logic of the "domain specificity" of cultural influence. As a political ideology, the governance function of the Long March Spirit may be more focused on shaping normative behaviors (e.g., discipline, collectivism) and non-financial objectives (e.g., ESG performance, crisis resilience) (Shen et al., 2025) [20] rather than directly driving short-term accounting profits. This aligns with North's (1990) [1] discussion on the scope of institutional constraints on behavior. Therefore, through comparative analysis, this study advocates that future cultural research should move beyond the assumption that "culture is a monolithic construct" and instead explore the differentiated governance pathways and boundary conditions of different cultural types.

5.2 Exploring the Mechanisms and Boundary Conditions of Heterogeneous Findings

The heterogeneity analysis reveals the boundary conditions under which the ICH effect operates, providing important contextualization for theory.

The stronger ICH effect in State-Owned Enterprises (SOEs) can be explained by theories of institutional isomorphism and resource complementarity. Due to their ownership nature, SOEs are more deeply embedded in local socio-political networks and have a natural affinity with ICH as a form of local social capital. This institutional fit allows SOEs to more effectively absorb and translate the trust networks and reputation mechanisms inherent in ICH (Dong et al., 2018) [16], leading to more pronounced performance improvements. This finding complements the existing literature on how ownership structure moderates cultural effects (Chen et al., 2019) [5], suggesting that in the presence of strongly embedded informal institutions like ICH, SOEs' "disadvantages" may transform into unique resource advantages.

The amplifying effect of older CEOs on the ICH effect resonates with the Upper Echelons Theory perspective on how managers' cognitive bases influence strategic choices. Older CEOs typically possess richer local social experience, deeper internalization of traditional values, and a stronger concern for long-term reputation. This makes them more adept at interpreting and utilizing the "artisan spirit" and sustainable development philosophy conveyed by ICH and translating them into conservative financial policies (reducing leverage). This finding extends cultural influence research from the macro-regional level to the micro-level of managerial characteristics, revealing a "human capital channel" for cultural value conversion.

5.3 Theoretical Significance of the Financial Conservatism Mediation Mechanism

The mechanism analysis confirms that reducing financial leverage is a key transmission channel through which ICH enhances performance. This finding has dual theoretical significance.

On one hand, it opens the "black box" of how culture influences corporate financial decision-making. Previous studies often stopped at correlational conclusions about "culture affecting performance," lacking tests of the intermediate economic decision-making processes. We demonstrate that the values of focus, prudence, and risk aversion advocated by ICH are concretely manifested in capital structure choices (Myers, 1977) [22], namely lower debt ratios. This provides a solid empirical link in the "culture-behavior-performance" chain.

On the other hand, it expands capital structure theory. Beyond traditional explanations based on trade-off theory, pecking order theory, and market perfection, our research shows that local informal cultural norms are an important yet overlooked factor influencing corporate leverage decisions. Firms do not make financial decisions in a vacuum; the "cultural air" they inhabit subtly shapes management's risk preferences and financing inclinations.

5.4 Synthesis of Research Contributions

In summary, the contributions of this study are multidimensional:

It systematically introduces ICH into corporate finance research for the first time and empirically distinguishes the different economic consequences of utilitarian traditional culture versus ideological revolutionary culture, enriching and refining the theoretical map of culture as an informal institution.

It pioneers the application of a DML framework combined with causal discovery in the field of cultural finance, providing a more rigorous causal inference solution for addressing the endogeneity inherent in cultural variables.

It not only validates the causal effect of ICH on performance but also reveals its contextual conditions (ownership, CEO age) and a specific financial transmission path (leverage reduction), making the findings more profound and actionable.

The study indicates that for firms, ICH is not merely an object to be protected but also a utilizable governance resource and strategic asset. For policymakers, differentiated strategies should be designed to harness the unique potential of different cultural types in promoting high-quality economic development.

7. Although limitations are acknowledged, the reflection on the core proxy variables and the underlying causal identification assumptions is not sufficiently in-depth. It is recommended that the limitations section be further elaborated and more clearly connected to future research directions.

6.2. Limitations and Avenues for Future Research

This study acknowledges several limitations, which also illuminate promising directions for future inquiry.

6.2.1 Limitations

Macro-level Variable Measurement. This study employs the number of Intangible Cultural Heritage (ICH) projects at the city level as a proxy for the regional cultural environment. While this approach effectively measures the concentration of cultural resources across regions, it cannot precisely capture the extent to which individual firms actively utilize, internalize, or strategically prioritize these ICH resources.

Challenges in Causal Identification. Although the Double Machine Learning method effectively controls for high-dimensional observable confounders, unobservable omitted variables—such as difficult-to-quantify local informal institutions, unique social capital, or tacit knowledge networks—may still pose a potential threat to the causal estimates.

Neglect of Spatial Dependence. This study assumes spatial independence in firm performance across the sample. However, in reality, factors like cultural diffusion, regional economic linkages, or policy imitation may lead to spatial autocorrelation in firm performance. Ignoring this spatial dependency structure could result in misestimated standard errors, thereby compromising the rigor of statistical inference.

Incomplete Mechanism Exploration. This study confirms financial conservatism (deleveraging) as one effective mediating pathway, but this is likely only one among multiple parallel mechanisms. The potential impacts of ICH on corporate brand value, stakeholder trust, and even craftsmanship-driven innovation have not been fully examined, leaving the complete picture of the transmission mechanisms to be fully mapped.

6.2.2 Avenues for Future Research

Building upon these limitations, future research can advance in the following directions:

Developing Micro-level Measurement Tools. Future studies could attempt to construct firm-level indicators of cultural embeddedness or utilization by employing textual analysis (mining references to specific cultural elements in annual reports or CSR reports), survey questionnaires (measuring the cultural values of management and employees), or case studies. This would deepen the measurement from "regional cultural environment" to "firm-specific cultural capital."

Expanding Causal Identification Strategies. To mitigate omitted variable bias, future research could seek cleaner exogenous policy shocks for identification, such as exploiting the quasi-natural experiment of a successful designation on the "Representative List of the Intangible Cultural Heritage of Humanity." Concurrently, future work should develop and apply a "Spatial Double Machine Learning" model. By incorporating a spatial weight matrix, this model would explicitly control for the spatial spillover effects of both performance and culture, thereby yielding more robust causal estimates while accommodating complex functional forms and high-dimensional controls.

Systematically Testing Multiple Mechanisms. It is recommended to adopt more systematic mediation analysis frameworks (e.g., multiple mediation models or causal mediation analysis) to concurrently test the various parallel pathways through which ICH might affect firm performance—such as "brand enhancement," "trust building," and "innovation incentive"—and to quantify the relative importance of each path.

 

Please contact me if you have any questions on the revised manuscript.

 

Best regards,

 

Fusheng Xie

 

Author Response File: Author Response.pdf