Corporate Social Responsibility and the Willingness to Eco-Innovate among Chilean Firms

Abstract

Adopting social responsibility can be a key strategy for firms to mitigate the impact of production on the environment, contributing to a more sustainable business model. Based on the triple bottom line perspective, we analyze the effect of Corporate Social Responsibility (CSR) on the willingness to allocate resources to eco-innovation among companies in a developing country. Firm-level data from the Fifth Longitudinal Survey of Chilean Firms are used to estimate a binary probit model for the willingness to eco-innovate and a Heckman sample-selection model for total expenditures and investment in eco-innovation. Results confirm that legal regulations and R&D efforts are drivers of eco-innovation among Chilean firms. Larger-sized and export-oriented firms also exhibit a higher willingness to eco-innovate. The main findings show a positive influence of CSR policy on the willingness to eco-innovate and on the resulting resource allocation decision. Interestingly, the evidence reveals that while financial and environmental CSR dimensions only affect the probability of adopting eco-innovation, the social CSR dimension also increases the amount firms spend on eco-innovation. This result suggests that social CSR may act as a complement to other CSR dimensions. These results are robust when controlling for firm-level specific effects under sample-selection models.

1. Introduction

In the last few decades, increasing pollution levels from human consumption and production activities have become a global concern for governments. From a production perspective, corporate social responsibility (CSR) can be an effective strategy to achieve long-term private economic success and mitigate the negative consequences of firm production on environmental quality and the capacity of nature to provide ecosystem services [1,2]. Changing corporate culture toward social responsibility also influences consumption patterns towards environmentally friendly products. This shift requires firms to adapt their production systems and corporate competitive strategies so that they focus on more sustainable business opportunities [3].

In that regard, entrepreneurial research and development (R&D) efforts have focused on new capabilities that facilitate the adoption of eco-innovation (also known as green or environmental innovation), minimizing environmental impact throughout product life cycles and establishing an ethical relationship with nature [4,5,6]. (Four types of eco-innovation have been highlighted in the literature: product, process, organizational, and marketing [7,8].) These efforts are in line with new challenges imposed by the Sustainable Development Goals (SDGs), where private companies have been called to play an active role in changing their production methods toward more sustainable business practices [9,10]. However, firms’ overall contribution to the SDGs is still limited and mostly a matter of seeking legitimacy [11]. As such, understanding firms’ progress in adopting eco-innovation practices is therefore critical to accelerating this productive transformation.

Research on the determinants and benefits of eco-innovation has mainly focused on developed countries—see, for instance, [5,12,13,14,15,16]. Among the benefits of eco-innovation, the literature highlights that firms can improve productivity and financial performance, despite the higher costs of pollution prevention. Based on the multidimensional nature of CSR, the recent literature explores the linkage between CSR and corporate eco-innovation activities using the triple bottom line perspective; the results indicate a positive influence of the financial and environmental dimensions of CSR on eco-innovation [15,17,18,19,20,21]. To our knowledge, the evidence for developing countries is scarce but growing. One exception is [22], who found that the environmental, financial, and social dimensions of CSR are positively related to eco-innovation, with eco-innovation serving a mediator role between CSR and Brazilian firms’ performance. This paper contributes to the literature with further empirical evidence regarding the main factors underlying eco-innovation decisions by firms in a developing setting. The context of a developing country is relevant for several reasons. First, since previous research on eco-innovation has been carried out mainly in Europe and the United States, more specific local and regional eco-innovation elements related to eco-innovation contexts and social processes have been neglected [23]. In this sense, results obtained for developed settings cannot be extrapolated to other contexts, since it is necessary to take into account the specific characteristics of each region/country. For example, the level of environmental awareness of its population or the development and features of its national innovation system can differ. Second, weaker environmental regulatory frameworks that characterize these contexts make voluntary CSR strategies more crucial for mitigating the impact of human activity on the environment. Third, eco-innovation practices are often not yet at commercial stages in developing countries, facing a series of issues such as increasing risks to investors, lack of available sources of venture capital, lack of knowledge of operations, and higher incremental costs [24]. Finally, firms in developing countries are involved in a global and highly competitive environment in which better-prepared firms from developed countries participate. Therefore, firms operating from a developing country need to accelerate their innovation processes in order to gain competitive advantages and to be able to face a highly competitive global market [25].

The empirical study is focused on Chile, an emerging Latin American economy that exhibits an upward trend in CO₂ emissions per capita, emissions that have grown twofold over the last thirty years [26]. In response, recent regulatory efforts have reinforced the protection of ecosystems in Chile. (For instance, Law No. 21,100, enacted in 2019, banned single-use plastic bags from stores, supermarkets, and food-delivery services, whereas Law No. 20,920, enacted in 2016, established a legal framework for waste management, extended producer responsibility, and recycling promotion among Chilean firms [27,28].) These legal changes have imposed several challenges on Chilean firms, leading them to modify their environmental practices and adapt their production systems in order to generate additional value through eco-innovation practices.

In particular, we use the triple bottom line perspective to analyze the effect of CSR on the willingness to spend on and invest in eco-innovation among Chilean firms. Firm-level data from the Fifth Longitudinal Survey of Chilean Firms are used to estimate a probit model indicating willingness to eco-innovate. Next, a Heckman sample-selection model is used to explain innovation intensity proxied by current expenditure and total investment on eco-innovation, respectively, and conditioned on whether a firm eco-innovates. Then, our data allow us to distinguish between the choice to eco-innovate as well as the intensity of eco-innovation, which is less common in the literature. The estimation strategy used in this study assumes that the factors underlying a firm’s choice to eco-innovate may differ from the factors explaining the intensity or magnitude of that eco-innovation. To our knowledge, there is no attempt in the literature to model the extent of eco-innovation efforts within our proposed empirical framework. In turn, the following research questions arise: (1) Are firms adopting CSR strategies more likely to carry out eco-innovation activities? (2) Are eco-innovation expenditures different between firms adopting and not adopting CSR strategies? (3) What is the role of CSR dimensions (i.e., environmental, financial, and social CSR) in promoting eco-innovation? (4) Is there a substitution or complementary effect between CSR dimensions?

Our results confirm the importance of legal regulations and R&D efforts as drivers of eco-innovation. In addition, larger-sized and export-oriented firms exhibit a higher willingness to eco-innovate. The main findings show a positive influence of CSR policy on the willingness to eco-innovate and on the resulting resource allocation decision. Interestingly, the evidence reveals that while financial and environmental CSR dimensions only affect the probability of adopting eco-innovation, the social CSR dimension increases the amount firms spend on eco-innovation. This result suggests that social CSR may act as a complement to other CSR dimensions. These results are robust when controlling for firm-level specific effects under sample-selection models.

2. Literature Review

2.1. Eco-Innovation

Advancing toward a higher level of economic development requires a strong commitment among firms to reduce their impact on the environment, which can be achieved by encouraging the adoption of eco-innovation practices [20,29]. In a rapidly changing environment, companies need to develop new capabilities that facilitate green innovation. The environmental benefits of green innovation have been highlighted in the literature. For instance, activities related to eco-innovation are helping to control long-term carbon emission levels in the United States [16]; in the United Kingdom, eco-innovation in the manufacturing industry has increased investment in environmental R&D [5]; eco-innovation and environmental policies are proven to have a positive and significant effect on reducing greenhouse gases among OECD countries [14]; and environmental taxes combined with initiatives to promote transitions to renewable energies have reduced carbon emissions, smog, particulate matter (i.e., PM 2.5), and greenhouse gases in China [30].

Empirical studies have used several indicators to measure eco-innovation as a dependent variable. As proposed by [31], eco-innovation can be measured through four approaches: input, immediate output, direct output, and indirect impact. Eco-innovation has been proxied by environmental research and development investment (input approach), environmental patents (immediate output approach), and value-added-to-emissions ratio (indirect output approach) [4,32]. Recently, [33] highlighted four types of eco-innovation based on firm-performance indicators: product innovation, process innovation, organization innovation, and marketing innovation.

The literature on eco-innovation has highlighted several drivers that determine companies’ decisions to eco-innovate, drivers that derive from organizational capabilities, environmental regulations, and demand-side requirements [34]. Regarding organizational capabilities, ref. [15] argued that financial resources are strongly linked to the level of eco-innovation achieved by Spanish firms, whereas ref. [12] found a positive relationship between environmental business culture and eco-innovation in Spain. Additionally, internal and external factors related to market competitiveness have a positive influence on eco-innovation among Chinese companies [4]. The evidence for developing countries revealed that export-oriented companies are more likely to eco-innovate in Taiwan [35] and that eco-innovation is enhanced through collaboration between different types of organizations—such as companies and universities—in Chile [25].

Concerning environmental regulations, ref. [13] showed that government regulation and environmental subsidies are key to successfully promoting eco-innovation processes in European firms. In [34], the authors highlighted that abatement costs had a positive impact on eco-innovation-related decisions made by manufacturing firms in the United Kingdom.

Related to demand-side requirements, ref. [36] argued that in Nigeria, customer demand and local market competitiveness determine a company’s decision of whether to adopt eco-innovation practices.

2.2. CSR and Its Relationship with Eco-Innovation

Recently, scholars have argued that CSR has emerged as a demand-side driver of environmental innovation practices among firms [37,38]. In turn, [39] emphasized the relevance of measuring CSR and its link to the innovation process, extending the concept to inclusive social responsibility. The authors provided a methodological framework to assess the above concept through a composite index that considers three dimensions: social, socially responsible attitude, and economic.

With regard to the relationship between CSR and eco-innovation, empirical research has highlighted a positive relationship between these two factors. In [18], the authors corroborated that strategic CSR has a positive influence on the likelihood of adopting innovative practices in products and processes among firms from Luxembourg. The evidence from two-step regressions also revealed a negative impact of CSR on firm growth due to the potential trade-off between responsive CSR and eco-innovation. Based on dynamic capability theory, ref. [40] found CSR as a positive driver of eco-innovation among Chinese manufacturing firms. After conducting a hierarchical regression analysis, the authors also found that CSR enhances dynamic capability, and both variables jointly reinforce the efforts on eco-innovative processes and products. In [41], the authors used the existence of CSR disclosure regulations as a quasi-natural experiment to assess the effect of CSR activities on eco-innovation. Findings from the staggered difference and differences approach revealed a positive effect of the adoption of CSR disclosure mandates on the number of environmental patents and citations among firms from treated countries for the period 2002–2017.

Given the multidimensional nature of CSR, these results are not informative enough to uncover the mechanisms underlying this relationship between CSR and eco-innovation. To fill that gap, empirical studies using the triple bottom line perspective (see, for instance, [20]) suggest that we must expand the definition of CSR into three dimensions: financial, social, and environmental. In [42], the authors added that CSR can be classified into three contexts: economic, political, and legal. Within this multidimensional understanding of CSR, [43] highlighted the role of the environmental dimension of CSR to drive innovation through product improvement, the invention of new types of technologies, and the generation of new practices. The literature also indicates that adopting environmental CSR is key to boosting eco-innovation in Pakistan and China. Pollution prevention seems to be the main eco-innovation mechanism in China, where a quadratic relationship between environmental CSR and sustainable environmental innovation has been found [20,44].

Researchers have argued that other CSR dimensions are also relevant in explaining eco-innovation. For example, [15] found that a company’s financial performance, amount of debt, available capital, and information traceability as measures of financial CSR are key factors that determine the level of eco-innovation adopted. In addition, ref. [45] argue that the social dimension of CSR—namely philanthropy and social integration—should translate into increased innovation practices within commercial and industrial sectors. In [19], the authors provided cross-country evidence that environmental and social CSR encourages innovation practices, results that are robust after addressing endogeneity issues. Using firm-level data from the Thomson Innovation database, the authors also found that the positive effect of environmental CSR is higher among firms located in developed countries.

We hypothesize that firms with a more consolidated CSR strategy are more likely to adopt eco-innovation practices and spend more resources on this type of action. We also expect that not only environmental CSR plays a role in boosting eco-innovation directly, but also other CSR strategies related to financial and social initiatives can be important to reinforce this effect. Then, exploring to what extent CSR dimensions are substitute or complementary in the promotion of eco-innovation emerges as a relevant research question to examine.

3. Materials and Methods

3.1. Empirical Strategy

Our dependent variable is the willingness to spend on and invest in eco-innovation among Chilean firms in 2017, which was proxied by current expenditure and total investment in environmental protection (questions C099 and C108, respectively, available in the Fifth Longitudinal Survey of Chilean Firms and known as ELE-5). Thus, we use an input approach to measure eco-innovation. In line with the ELE-5 questionnaire, both proxy variables are built on firms’ self-reported information about the number of financial resources they allocated to a series of pollution mitigation efforts, such as air and climate protection, wastewater treatment, waste management, soil and water resources protection, noise pollution reduction, and others. In order to model eco-innovation decisions, we first estimated the following binary probit model:

$$P(y_i=1| C,X,W,Z)=\mathsf{\Phi }\left(C\varphi +X\beta +W\gamma +Z\theta \right)$$

(1)

where $y_i$ is a dummy variable that takes the value of one if the reported current expenditure (or total investment) on eco-innovation spent by firm $i$ was greater than zero in 2017, and zero otherwise. Because we are interested in addressing the influence of CSR dimensions on eco-innovation-related decisions made by Chilean firms, we define $C$ as a vector of explanatory variables linked to corporate social responsibility, which reflects a CSR intensity index. Given the data available on the ELE-5 survey, this index ranges from zero to three: 0 if firm $i$ has no CSR policy, 1 if firm $i$ has a CSR policy (question H023), 2 if firm $i$ computes CSR indicators (question H024), and 3 if firm $i$ publicly publishes the above indicators (question H026). (Note that this index constitutes a pseudo-measure of CSR transparency and accountability; see [46,47] for a discussion regarding the above attributes of CSR and their effect on business performance.) We expect that firms with a larger CSR index, suggesting a more consolidated CSR strategy, will be more likely to adopt eco-innovation-related actions. Following the triple bottom line perspective [20], this index was also computed for environmental and social CSR dimensions, which were averaged from several CSR sub-dimensions. These statistics are listed in

Table A1. Sub-dimensions used to calculate the environmental and social CSR intensity index.

	ELE-5 Survey Questions:
CSR Dimensions and Sub-Dimensions	There Is a CSR Policy (Yes = 1)	CSR Indicators Are Computed (Yes = 1)	CSR Indicators Are Published (Yes = 1)
Social CSR policies on:
Diversity and inclusion	H028	H029	H031
Ethics code	H033	H034	H036
Gender	H038	H039	H041
Inclusion of people with disabilities	H043	H044	H046
Community engagement	H068	H069	H071
Environmental CSR policies on:
Energy efficiency	H048	H049	H051
Waste management	H053	H054	H056
Carbon footprint	H058	H059	H061
Water footprint	H063	H064	H066

Source: Authors’ own elaboration based on the ELE-5 survey questionnaire.

. Financial CSR intensity was calculated as the average between two dummy variables that take the value of one if banks are the company’s main source of financing (question B015) and if the firm made use of enterprise-resource-planning (ERP) software (question J076), and zero otherwise. We also expect a positive effect of each CSR dimension, as included separately, on the adoption of eco-innovation practices.

$X$ is a vector of covariates related to the firms’ financial performance in 2017. Specifically, return on assets (ROA) was computed as the total earnings (question C041) to assets (question C062) ratio, whereas financial leverage was defined by a total liabilities (question C072) to assets ratio.

$W$ is a vector of regressors related to the firms’ business environment and growth factors, such as the importance of competition (question H084) and legal regulations (question H085), and the use of social media for customer engagement in product/services innovation (question J055).

$Z$ is a vector of firm characteristics that include firm size (question TAMANO), board of directors’ size (questions H014 and H015), dummy variables for firms’ export orientation (questions D178 and D179), the use of social media for marketing image and products (question J051), R&D efforts (questions H072, H073, and H074), and economic sector (question GLOSA_CIIU).

Note that financial indicators included in vector $X$ and covariates comprised in vectors $W$ and $Z$ have been frequently used as control variables in related literature, e.g., [19,20,40,41]. For illustrative and replication purposes,

Table A2. Variables description.

Variables	Description
Dependent variables
Willingness to eco-innovate (expenditure)	1 if C099 > 0; 0 otherwise
Willingness to eco-innovate (investment)	1 if C108 > 0; 0 otherwise
Current expenditure on eco-innovation	log(1 + C099)
Total investment in eco-innovation	log(1 + C108)
Explanatory variables
CSR policy and dimensions:
CSR policy (intensity index)	H023 + H024 + H026
Environmental CSR	[(H048 + H049 + H051) + (H053 + H054 + H056) + (H058 + H059 + H061) + (H063 + H064 + H066)]/4
Financial CSR	(B015 + J076)/2
Social CSR	[(H028 + H029 + H031) + (H033 + H034 + H036) + (H038 + H039 + H041) + (H043 + H044 + H046) + (H068 + H069 + H071)]/5
Firm’s financial performance:
ROA	C041/C062
Leverage	C072/C062
Business environment:
Importance of competition	4–H084
Importance of legal regulations	4–H085
Customer engagement in product innovation	1 if J055 = 1; 0 otherwise
Firm characteristics:
Firm size	6–TAMANO
Board of directors’ size	H014 + H015
Firm’s export orientation	1 if D178 = 1 or D179 = 1; 0 otherwise
Social media for marketing image	1 if J051 = 1; 0 otherwise
R&D efforts	1 if (H072 + H073 + H074) > 0; 0 otherwise

Source: Authors’ own elaboration based on data from the ELE-5 survey.

describes how the dependent variables and covariates were computed using available data from the ELE-5 survey (see Appendix A). Finally, $\varphi$, $\beta$, $\gamma$, and $\theta$ are the vectors of parameters to be estimated, and $\mathsf{\Phi }$ is the normal cumulative density function.

Secondly, since eco-innovation practices are in an incipient stage in Chile (i.e., most firms are not devoting any financial efforts to these actions), we estimated a Tobit model in order to address the censored lower limit observed in current expenditure and total investment allocated to eco-innovation, which is formalized as follows:

$$y_i=\left\{\begin{array}{cc}0;\hfill & \mathrm{if} y_i^{\ast }\le 0\hfill \\ y_i^{\ast };\hfill & \mathrm{if} y_i^{\ast }>0\hfill \end{array}\right.$$

(2)

$$y_i^{\ast }=C_i\varphi +X_i\beta +W_i\gamma +Z_i\theta +{\epsilon }_i$$

(3)

where $y_i$ is the dependent variable for the natural logarithm of one plus current expenditure (or total investment) on eco-innovation in Chilean pesos, whereas $y_i^{\ast }$ is an unobserved latent variable. Vectors $C$, $X$, $W$, and $Z$ are defined as above, $\varphi$, $\beta$, $\gamma$, and $\theta$ are the vectors of parameters to be estimated, and ${\epsilon }_i$ is the estimation error, which is distributed $\mathcal{N}\left(0, {\sigma }^2\right)$. We expect that firms adopting more advanced CSR strategies will spend and invest more in eco-innovation projects, especially those implementing environmental and financial CSR strategies.

Alternatively, we assumed that the eco-innovation decision follows a different process than the resource allocation decision in terms of how much a firm spends on or invests in eco-innovation. Therefore, we estimated the following Heckman sample-selection model [48,49]:

$$s_i=\left\{\begin{array}{cc}0;\hfill & \mathrm{if} s_i^{\ast }\le 0\hfill \\ 1;\hfill & \mathrm{if} s_i^{\ast }>0\hfill \end{array}\right.$$

(4)

$$y_i=\left\{\begin{array}{cc}0;\hfill & \mathrm{if} s_i^{\ast }\le 0\hfill \\ y_i^{\ast };\hfill & \mathrm{if} s_i^{\ast }>0\hfill \end{array}\right.$$

(5)

$$s_i^{\ast }=C_i{\varphi }_1+X_i{\beta }_1+W_i{\gamma }_1+Z_i{\theta }_1+{\epsilon }_{1i}$$

(6)

$$y_i^{\ast }=C_i{\varphi }_2+X_i{\beta }_2+W_i{\gamma }_2+Z_i{\theta }_2+{\epsilon }_{2i}$$

(7)

where $s_i$ is the selection process for the willingness to eco-innovate, i.e., $s_i=1$ if question C099 (or C108) is greater than zero, and zero otherwise, $y_i$ is the natural logarithm of one plus current expenditure (or total investment) on eco-innovation, and $s_i^{\ast }$ and $y_i^{\ast }$ are unobserved latent variables. Vectors $C$, $X$, $W$, and $Z$ are defined as above, whereas $\varphi$, $\beta$, $\gamma$, and $\theta$ are the vectors of parameters to be estimated. The error terms ${\epsilon }_{1i}$ and ${\epsilon }_{2i}$ follow a bivariate normal distribution, where their covariance and correlation are equal to ${\sigma }_{12}$ and $\rho$, respectively; thus, the inverse hyperbolic tangent of $\rho$ (athrho) was also estimated. We expect that factors underlying the probability of eco-innovation will differ from those explaining the magnitude or intensity of the eco-innovation measured by expenditures or investment. In addition, it is expected that while the environmental dimension of CSR affects both the willingness and intensity to eco-innovate, financial and social dimensions have differentiated impacts.

3.2. Data

Our empirical study used data from the ELE-5 survey conducted in 2018 by the Chilean Ministry of Economy and the National Institute of Statistics (INE), which gathered firm-level data regarding companies’ characteristics and their business environment for the years 2016 and 2017. Its questionnaire—which included binary, nominal, ordinal, and open queries—was divided into five sections: accounting and finance; markets, customers, and suppliers; general management; human resources; and information and communication technologies. Although the ELE-5 survey excluded firms reporting annual sales below 21.4 million Chilean pesos (or 52,000 international US dollars), its weighted sample is representative of Chilean firms at the firm size and sectoral levels. The response rate for the ELE-5 survey was 53.7% and most of the respondents belong to the commerce (34.3%), manufacturing industry (10.6%), transportation (10.1%), and building (9.5%) sectors [50,51].

Descriptive statistics for dependent variables and covariates are reported in

Table 1. Descriptive statistics.

	N	Mean	Std. Dev.	Min.	Max.
Dependent variables
Willingness to eco-innovate (expenditure)	6075	0.0693	0.254	0	1
Willingness to eco-innovate (investment)	5879	0.0180	0.133	0	1
Current expenditure on eco-innovation	6075	0.6657	2.566	0	18.7303
Total investment in eco-innovation	6075	0.1821	1.450	0	19.9881
Explanatory variables
CSR policy and dimensions:
CSR policy (intensity index)	6075	0.3990	0.791	0	3
Environmental CSR	6075	0.2300	0.528	0	3
Financial CSR	6075	0.4699	0.378	0	1
Social CSR	6075	0.3173	0.554	0	3
Firm’s financial performance:
ROA	6075	2.96 × 106	2.3 × 108	0	1.79 × 1010
Leverage	6075	11.2699	590.037	−27.7678	41,250
Business environment:
Importance of competition	6075	2.3384	0.770	1	3
Importance of legal regulations	6075	2.0013	0.819	1	3
Customer engagement in product innovation	6075	0.1370	0.344	0	1
Firm characteristics:
Firm size	6075	3.5251	1.527	1	5
Board of directors’ size	6075	1.4532	5.660	0	315
Firm’s export orientation	6075	0.1221	0.327	0	1
Social media for marketing image	6075	0.2856	0.452	0	1
R&D efforts	6075	0.2016	0.401	0	1

. Figures from Table 1 suggest a very low level of eco-innovation in Chile. Only 6.9% and 1.8% of firms report having spent or invested resources on eco-innovation, respectively. These figures are lower than those reported in the Chilean survey of Enterprise Innovation (known as ENI), which suggests that 25% of Chilean firms adopt some type of eco-innovation [25]. The reason behind this discrepancy is that this paper focuses on innovations devoted to environmental damage reduction and health and safety improvement actions (i.e., pollution mitigation efforts), which tend to underestimate the level of adoption of eco-innovation practices.

On average, firms spent around USD 150,000 in mitigating pollution and protecting the environment in 2017. This finding confirms the urgency and need for intervention in order to accelerate the process of eco-innovation adoption among Chilean firms. Complementary figures also suggest a low adoption of CSR policies in Chile, reported by only 25.8% of firms (see

Table A3. Indicators regarding CSR policies implemented by Chilean firms by enterprise size and CSR dimensions.

	All Firms	Firm Size Classification:
		Micro	Small	Medium	Large
Number of firms	6075	1055	1595	909	2516
Share of Chilean firms with a CSR policy in force:
CSR policy	25.8%	9.1%	16.2%	24.4%	39.3%
Environmental CSR	28.8%	11.4%	17.4%	31.1%	42.6%
Financial CSR	68.2%	36.1%	55.2%	76.5%	87.0%
Social CSR	44.7%	18.3%	26.7%	43.6%	67.6%
Average CSR intensity score:
CSR policy	0.399	0.134	0.212	0.338	0.651
Environmental CSR	0.230	0.073	0.108	0.209	0.381
Financial CSR	0.470	0.206	0.318	0.530	0.655
Social CSR	0.317	0.102	0.153	0.270	0.529

Source: Authors’ own elaboration based on data from the ELE-5 survey.

, Appendix A). Fewer firms report incorporating CSR initiatives within the environmental dimension as compared to financial and social CSR dimensions. As such, there is a prominent gap in the implementation of a CSR policy, particularly in the environmental dimension, among micro- and small-sized firms when compared to large-sized enterprises. This gap is also observable in the amount of transparency and accountability of CSR in differently sized Chilean firms (see Table A3, Appendix A).

4. Results and Discussion

4.1. Regarding the Willingness to Eco-Innovate among Chilean Firms

Table 2. Average marginal effects after probit estimation for the willingness to eco-innovate.

Dependent Variable:	Willingness to Eco-Innovate (Yes = 1)
	Based on Current Expenditure			Based on Total Investment
	(1)	(2)	(3)	(4)	(5)	(6)
CSR policy and dimensions:
CSR policy (intensity index)	0.014 ***			0.008 ***
	(0.00)			(0.00)
Environmental CSR		0.043 ***	0.051 ***		0.012 ***	0.012 ***
		(0.00)	(0.01)		(0.00)	(0.00)
Financial CSR			0.031 ***			0.012 *
			(0.01)			(0.01)
Social CSR			−0.014 *			0.001
			(0.01)			(0.00)
Firm’s financial performance:
ROA	2.5 × 10−11 ***	2.3 × 10−11 ***	2.3 × 10−11 ***	−0.000	−0.000	−1.5 × 10−9 *
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Leverage	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Business environment:
Importance of competition	−0.006	−0.007	−0.007	−0.003	−0.003	−0.003
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Importance of legal regulations	0.017 ***	0.015 ***	0.015 ***	0.006 *	0.005 *	0.005 *
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Customer engagement	−0.004	−0.004	−0.003	0.002	0.003	0.002
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Firm characteristics:
Firm size	0.028 ***	0.024 ***	0.021 ***	0.003	0.003	0.001
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Board of directors’ size	0.002 **	0.002 *	0.002 *	0.0005 ***	0.0004 **	0.0004 **
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Firm’s export orientation	0.024 *	0.017	0.015	0.005	0.004	0.003
	(0.01)	(0.01)	(0.01)	(0.00)	(0.00)	(0.00)
Social media for marketing image	0.009	0.004	0.003	0.005	0.004	0.003
	(0.01)	(0.01)	(0.01)	(0.00)	(0.00)	(0.00)
R&D efforts	0.041 ***	0.036 ***	0.034 ***	0.016 ***	0.015 **	0.014 **
	(0.01)	(0.01)	(0.01)	(0.00)	(0.00)	(0.00)
Observations	6075	6075	6075	5879	5879	5879
Log likelihood	−1268.8	−1231.3	−1223.1	−413.2	−409.1	−406.4
Pseudo $R^2$	0.171	0.195	0.200	0.221	0.229	0.234
${\chi }^2$ statistic	396.0	471.8	478.7	212.8	236.4	232.8

Note: Sectorial dummies are unreported. Robust standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.

reports the average marginal effects (AMEs) estimated by the binary probit model for the willingness to eco-innovate. The evidence indicates that the greater intensity of CSR policy, the higher the willingness to eco-innovate will be among Chilean firms, validating our hypothesis at the 0.1% level (columns 1 and 4, Table 2). This means that firms with a more complex implementation of CSR policy are more likely to take action to reduce the impact of their operations on the environment. Although we used the input approach to proxy eco-innovation, this finding is in line with recent research in which dependent variables are based on the immediate output approach [19,41].

Focusing on the multidimensional character of CSR, the results confirm that the environmental dimension is positively related to the willingness to eco-innovate (columns 2–3 and 5–6, Table 2), which has been previously reported in the literature [19,20]. This result is reasonable since adopting CSR in the environmental dimension comes with a commitment to implement tangible and low-environmental-impact corporate actions, which need resources to materialize. In addition, the evidence suggests that as firms commit more to the financial dimension of CSR, they increase their inclination to eco-innovate (columns 3 and 6, Table 2). CSR policies in the financial dimension may signal that firms have better access to and planning of financial resources, which may facilitate the financing of eco-innovative actions. On the contrary, CSR policies in the social dimension result in a reduced willingness to eco-innovate among Chilean firms when this decision is based on current expenditure (column 3, Table 2). This negative effect on the eco-innovation of social CSR policies may indicate a sort of substitution effect between social and environmental dimensions of CSR. In contrast, the financial dimension of CSR is complementary to eco-innovation because resources are needed to finance initiatives with lower environmental impact.

With regard to firms’ financial performance, we found a positive and statistically significant relationship between the ROA ratio and the willingness to spend on eco-innovations (columns 1–3, Table 2). This finding confirms the importance of financial performance in boosting the willingness to devote resources to eco-innovation. That said, the ROA ratio reduces the willingness to eco-innovate when based on total investment (column 6, Table 2). This result contradicts previous cross-country evidence regarding the effect of the ROA ratio on eco-innovation when the latter has been measured according to the immediate output approach, e.g., [41].

Regarding the effect of the business environment on the decision to eco-innovate, we can identify legal regulations as a key growth factor and positive driver of green innovation among Chilean firms (columns 1–6, Table 2).

The evidence shows that eco-innovation is more likely among larger-sized firms (columns 1–3, Table 2), which is consistent with the evidence reported by [18,19,20,42]. A plausible explanation is that larger companies have higher levels of available resources and a solid organizational structure to invest in and promote technologically innovative practices.

The results show a higher inclination to eco-innovate as the board of directors’ size increases (columns 1–6, Table 2), which has been previously indicated by [20]. R&D efforts also seem to exacerbate the willingness of firms to eco-innovate (columns 1–6, Table 2). In this regard, [13] remarked that applied R&D provides technological resources and internal capabilities, two factors that are crucial for encouraging eco-innovation practices. The results also suggest that the likelihood of eco-innovation is higher among export-oriented firms (column 1, Table 2), a finding that has been previously reported for developing countries, e.g., [35]. Finally, no evidence was found for the effect of competition as a growth factor or for the use of social media for customer engagement and marketing image/products on the willingness to adopt green innovations.

4.2. Regarding the Resource Allocation Decision to Eco-Innovation

We were interested in how Chilean firms make resource allocation decisions in terms of expenditure and investment in environmental innovation. Given that most firms report not spending on or investing in eco-innovation, we have to deal with a large number of observations at the zero-accumulation point. The Tobit model does well in accounting for this in estimations.

Table 3. Average marginal effects after Tobit estimation for resource allocation to eco-innovation.

Dependent Variable:	Current Expenditure on Eco-Innovation			Total Investment in Eco-Innovation
	(1)	(2)	(3)	(4)	(5)	(6)
CSR policy and dimensions:
CSR policy (intensity index)	0.156 ***			0.082 ***
	(0.03)			(0.02)
Environmental CSR		0.423 ***	0.471 ***		0.124 ***	0.113 ***
		(0.04)	(0.06)		(0.02)	(0.03)
Financial CSR			0.306 ***			0.124 *
			(0.09)			(0.05)
Social CSR			−0.083			0.018
			(0.06)			(0.03)
Firm’s financial performance:
ROA	5.2 × 10−11 **	5.7 × 10−11 ***	5.6 × 10−11 ***	−0.024	−0.026	−1.5 × 10−8 *
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Leverage	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Business environment:
Importance of competition	−0.066	−0.071	−0.075	−0.030	−0.032	−0.032
	(0.04)	(0.04)	(0.04)	(0.02)	(0.02)	(0.02)
Importance of legal regulations	0.171 ***	0.148 ***	0.151 ***	0.055 *	0.052 *	0.049 *
	(0.04)	(0.04)	(0.04)	(0.02)	(0.02)	(0.02)
Customer engagement	−0.031	−0.018	−0.014	0.023	0.034	0.027
	(0.09)	(0.09)	(0.09)	(0.05)	(0.05)	(0.05)
Firm characteristics:
Firm size	0.296 ***	0.261 ***	0.226 ***	0.036 *	0.034 *	0.017
	(0.03)	(0.03)	(0.03)	(0.02)	(0.02)	(0.02)
Board of directors’ size	0.010 ***	0.007 **	0.007 *	0.004 ***	0.004 ***	0.004 ***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Firm’s export orientation	0.250 **	0.178 *	0.159	0.054	0.042	0.037
	(0.09)	(0.09)	(0.09)	(0.05)	(0.05)	(0.04)
Social media for marketing image	0.113	0.065	0.051	0.048	0.042	0.035
	(0.08)	(0.08)	(0.07)	(0.04)	(0.04)	(0.04)
R&D efforts	0.383 ***	0.337 ***	0.310 ***	0.150 ***	0.142 **	0.131 **
	(0.08)	(0.08)	(0.08)	(0.04)	(0.04)	(0.04)
Observations	6075	6075	6075	6074	6074	6074
Log likelihood	−2591.5	−2553.6	−2546.9	−755.9	−752.1	−749.0
Pseudo $R^2$	0.097	0.110	0.112	0.144	0.148	0.152
$F$ statistic	27.8	37.3	35.4	17.1	26.3	23.2

Note: Sectorial dummies are unreported. Robust standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.

compiles the estimated AMEs after a Tobit estimation for the current expenditure and total investment allocated to eco-innovation by Chilean firms, respectively.

The estimates from the Tobit model confirm the positive influence of CSR policy and its environmental dimension on the number of resources allocated to eco-innovation (columns 1–6, Table 3). Note that the magnitude of the effect of CSR on eco-investment is relatively lower than that found on eco-expenditures. In other words, long-term decisions linked to investment in eco-innovation are less sensitive to CSR intensity than short-term expenditure decisions. In addition, we found that CSR policy in the social dimension is statistically insignificant in explaining investment and expenditure in eco-innovation, although CSR in the financial dimension seems to impact the use of resources and investment in eco-innovation (columns 3 and 6, Table 3). Thus, firms with better financial strategies have larger expenditures in eco-innovation, confirming the importance of good financial health among firms as a precursor to devoting more resources to eco-innovation.

In exploring the effect of firms’ financial performance, the positive and significant effect of the ROA ratio confirms the importance of a firm’s sound financial situation on its eco-innovation expenditure (columns 1–3, Table 3). As previously indicated, results show that both spending on and investing in eco-innovation are greater among larger-sized firms (columns 1–5, Table 3). In Chilean firms, the importance of legal regulations on firms’ growth, the board of directors’ size, and R&D practices explain the amount of financial resources devoted to environmental innovation (columns 1–6, Table 3). Additionally, current expenditure on eco-innovation seems to be positively influenced by a firm’s export orientation (columns 1–2, Table 3).

Furthermore, and as previously highlighted, there is no statistically significant evidence to support the effect of competition as a growth factor, or that the use of social media for customer engagement and marketing image/products impacts a firm’s decision to allocate resources to eco-innovation.

The decision to eco-innovate may differ from the decision to expand or alter the degree of eco-innovation. To explore this further, we estimate a Heckman sample-selection model to explain the amount firms invest in or spend on eco-innovation while accounting for the fact that eco-innovating firms are not a random sample of the population. This model allows us to split our specification into two decisions: one predicting whether firms adopt eco-innovation, and another explaining the amount of investment or expenditure for those that have adopted.

Table 4. Estimates from the Heckman model with sample selection of current expenditure on eco-innovation.

	Selection Model	Current Expenditure	Selection Model	Current Expenditure	Selection Model	Current Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
CSR policy and dimensions:
CSR policy (intensity index)	0.125 ***	0.645 ***
	(0.03)	(0.14)
Environmental CSR			0.396 ***	0.321	0.479 ***	−0.494
			(0.04)	(0.28)	(0.06)	(0.37)
Financial CSR					0.298 ***	−0.044
					(0.08)	(0.44)
Social CSR					−0.132 *	1.364 ***
					(0.06)	(0.27)
Firm’s financial performance:
ROA	2.0 × 10−10 ***	−0.000	1.9 × 10−10 ***	0.000	1.8 × 10−10 ***	−0.000
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Leverage	−0.000	−0.004	−0.000	−0.001	−0.000	−0.007
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Business environment:
Importance of competition	−0.055	−0.110	−0.062	−0.114	−0.068	−0.073
	(0.04)	(0.17)	(0.04)	(0.18)	(0.04)	(0.17)
Importance of legal regulations	0.153 ***	0.048	0.137 ***	0.037	0.142 ***	−0.057
	(0.04)	(0.20)	(0.04)	(0.19)	(0.04)	(0.18)
Customer engagement	−0.038	0.447	−0.036	0.755	−0.030	0.345
	(0.09)	(0.41)	(0.09)	(0.42)	(0.09)	(0.40)
Firm characteristics:
Firm size	0.248 ***	0.825 ***	0.223 ***	0.793 ***	0.191 ***	0.729 ***
	(0.03)	(0.22)	(0.03)	(0.18)	(0.03)	(0.18)
Board of directors’ size	0.018 *	−0.014	0.016 *	−0.012 *	0.017 *	−0.010
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Firm’s export orientation	0.198 **	0.365	0.152 *	0.279	0.139	0.348
	(0.07)	(0.33)	(0.07)	(0.32)	(0.08)	(0.31)
Social media for marketing image	0.074	0.727 *	0.034	0.637	0.023	0.614
	(0.07)	(0.33)	(0.07)	(0.33)	(0.07)	(0.33)
R&D efforts	0.327 ***	−0.527	0.302 ***	−0.492	0.286 ***	−0.528
	(0.06)	(0.34)	(0.06)	(0.31)	(0.06)	(0.32)
Constant	−2.917 ***	6.194 *	−2.853 ***	6.806 **	−2.815 ***	7.093 **
	(0.24)	(2.84)	(0.24)	(2.22)	(0.24)	(2.40)
Inverse of Mills’ ratio	−1.091		−1.333		−1.429
	(0.89)		(0.69)		(0.77)
Observations	6075		6075		6075
Log likelihood	−2227.1		−2200.0		−2175.9

Note: Sectorial dummies are unreported. Robust standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.

and

Table 5. Estimates from the Heckman model with sample selection of total investment in eco-innovation.

	Selection Model	Current Expenditure	Selection Model	Current Expenditure	Selection Model	Current Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
CSR policy and dimensions:
CSR policy (intensity index)	0.224 ***	1.028 ***
	(0.05)	(0.27)
Environmental CSR			0.342 ***	0.806	0.322 ***	0.257
			(0.05)	(0.49)	(0.08)	(0.51)
Financial CSR					0.342 *	0.778
					(0.14)	(1.03)
Social CSR					0.037	1.232 **
					(0.10)	(0.46)
Firm’s financial performance:
ROA	−0.000	0.000	−0.000	0.000	−4.4 × 10−8 *	0.000
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Leverage	−0.000	−0.003	−0.000	0.001	−0.000	−0.008 *
	(0.00)	(0.01)	(0.00)	(0.00)	(0.00)	(0.00)
Business environment:
Importance of competition	−0.080	0.105	−0.077	−0.505	−0.079	−0.349
	(0.06)	(0.41)	(0.06)	(0.41)	(0.06)	(0.40)
Importance of legal regulations	0.155 **	−0.054	0.147 *	−0.384	0.141 *	−0.578
	(0.06)	(0.45)	(0.06)	(0.46)	(0.06)	(0.45)
Customer engagement	0.056	1.574	0.088	1.186	0.077	0.561
	(0.13)	(0.82)	(0.13)	(0.85)	(0.13)	(0.78)
Firm characteristics:
Firm size	0.090	1.055 ***	0.086	0.947 ***	0.042	0.694 *
	(0.05)	(0.20)	(0.05)	(0.25)	(0.05)	(0.27)
Board of directors’ size	0.013 ***	−0.059 *	0.011 ***	−0.006	0.011 ***	−0.006
	(0.00)	(0.02)	(0.00)	(0.01)	(0.00)	(0.01)
Firm’s export orientation	0.135	0.567	0.112	0.146	0.102	0.304
	(0.12)	(0.80)	(0.12)	(0.73)	(0.12)	(0.70)
Social media for marketing image	0.126	−0.112	0.108	−0.302	0.093	−0.336
	(0.12)	(0.60)	(0.12)	(0.70)	(0.12)	(0.68)
R&D efforts	0.385 ***	−0.346	0.370 ***	−0.825	0.349 ***	−0.909
	(0.10)	(0.59)	(0.10)	(0.67)	(0.10)	(0.63)
Constant	−3.370 ***	3.190	−3.511 ***	11.486 ***	−3.475 ***	12.021 ***
	(0.42)	(3.01)	(0.38)	(2.90)	(0.38)	(2.90)
Inverse of Mills’ ratio	−0.530		−1.821 *		−1.926 *
	(0.57)		(0.82)		(0.86)
Observations	6075		6075		6075
Log likelihood	−657.7		−662.3		−654.1

Note: Sectorial dummies are unreported. Robust standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.

report the estimates from the Heckman sample-selection model for the willingness to allocate resources to eco-innovation activities.

The complementary evidence confirms the positive effect of CSR policy on both the willingness to eco-innovate and the resource allocation decision, i.e., expenditure and investment (columns 1–2, Table 4 and Table 5). Interestingly, our findings reveal a dissimilar effect of CSR dimensions on firms’ inclination to eco-innovate. Results show that environmental and financial CSR only influence the willingness to eco-innovate but not the number of resources devoted to eco-innovation (columns 3–5, Table 4 and Table 5). In addition, current expenditure on eco-innovation only increases if the social dimension of CSR indicates a strong commitment to transparency among stakeholders and the community, although this characteristic of social CSR seems to reduce the willingness of firms to eco-innovate in the first place (columns 3–4, Table 5). Note that this finding is in line with cross-country studies that have found that corporate innovation is boosted by social CSR [19]. In other words, social CSR becomes complementary in boosting both expenditure and investment in eco-innovation once the decision to eco-innovate is already made. In turn, both the willingness to invest in eco-innovation and the resources involved in that decision are enhanced by a more transparent environmental and financial CSR (columns 3–4,

Table 6. Estimates from random-effects model of current expenditure and total investment in eco-innovation.

Dependent Variable:	Current Expenditure on Eco-Innovation			Total Investment in Eco-Innovation
	(1)	(2)	(3)	(4)	(5)	(6)
CSR policy and dimensions:
CSR policy (intensity index)	0.065 ***			0.072 ***
	(0.02)			(0.02)
Environmental CSR		0.103 ***	0.156 **		0.076 ***	0.122 **
		(0.02)	(0.05)		(0.02)	(0.05)
Financial CSR			0.052			0.091 **
			(0.04)			(0.03)
Social CSR			−0.065			−0.058
			(0.05)			(0.04)
Firm’s financial performance:
ROA	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Leverage	−0.002	−0.002	−0.002	−0.002	−0.002	−0.002
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Business environment:
Importance of competition	−0.010	−0.012	−0.012	−0.063 **	−0.064 ***	−0.065 ***
	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)
Importance of legal regulations	0.020	0.019	0.019	0.026	0.026	0.025
	(0.02)	(0.02)	(0.02)	(0.01)	(0.01)	(0.01)
Customer engagement	0.000	0.000	0.001	0.076	0.079	0.080
	(0.04)	(0.04)	(0.04)	(0.05)	(0.05)	(0.05)
Firm characteristics:
Firm size	0.175 ***	0.178 ***	0.177 ***	0.058 ***	0.061 ***	0.054 ***
	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)
Board of directors’ size	0.012 *	0.012 *	0.012 *	0.011 ***	0.011 ***	0.011 ***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Firm’s export orientation	0.141 **	0.144 **	0.141 **	0.118 *	0.120 *	0.115 *
	(0.05)	(0.05)	(0.05)	(0.05)	(0.05)	(0.05)
Social media for marketing image	0.034	0.038	0.037	−0.016	−0.011	−0.015
	(0.05)	(0.05)	(0.05)	(0.03)	(0.03)	(0.03)
R&D efforts	0.062	0.059	0.058	0.102 **	0.103 **	0.100 **
	(0.05)	(0.05)	(0.05)	(0.03)	(0.03)	(0.03)
Constant	−0.150	−0.168	−0.182	−0.063	−0.072	−0.085
	(0.21)	(0.21)	(0.21)	(0.09)	(0.09)	(0.09)
Observations	8004	8004	8004	8004	8004	8004
Number of firms	4002	4002	4002	4002	4002	4002
Overall $R^2$	0.092	0.097	0.098	0.085	0.086	0.087
${\chi }^2$ statistic	218.3	221.2	223.6	105.7	106.6	106.5

Note: Sectorial dummies and time effects are unreported. Robust standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.

). Together, these results mean that among eco-innovators, social CSR reinforces the effect of environmental CSR on eco-innovation rather than competing with other CSR dimensions.

Concerning firms’ financial performance, our evidence shows that the effect of financial performance on eco-innovation-related decisions differs depending on financial performance indicators and the variable used to proxy the number of resources devoted to environmental protection activities. That is, while a higher ROA increases the willingness to spend on eco-innovation (columns 1, 3, and 5, Table 4), a higher leverage ratio appears to negatively influence the amount invested in environmental protection (column 6, Table 5). This last result indicates that larger debts and obligations relative to capital assets undermine additional financial efforts in pro-environmental practices, a finding that has been previously highlighted in the literature on eco-innovation [41].

Regarding the business environment, our results confirm that the willingness to eco-innovate is positively affected by the higher importance of legal regulations on firm performance (columns 1, 3, and 5, Table 4 and Table 5). This result indicates that firms embedded in a highly competitive industry, facing more demanding legal regulations, seem to prioritize reducing the impact of production on the environment, especially in contexts where eco-innovation is still an emerging corporate strategy.

In reference to firms’ characteristics, estimates from the Heckman selection model indicate that larger-sized firms show a higher proclivity to spend on and invest in eco-innovation (columns 1–6, Table 4; columns 2, 4, and 6, Table 5). The evidence also indicates that the willingness to eco-innovate is positively influenced by the board of directors’ size (columns 1, 3, and 5, Table 4 and Table 5); however, the resources allotted to eco-innovation shrink as the number of board members grows (column 4, Table 4; column 2, Table 5). In other words, once firms commit to eco-innovating, a smaller board may facilitate agreements to invest more resources into eco-innovation. R&D efforts encourage the willingness to spend on and invest in eco-innovation (columns 1, 3, and 5, Table 4 and Table 5), and a firm’s export orientation is also a positive driver for implementing actions with lower environmental impact (columns 1 and 3, Table 4). Our finding on the effect of R&D may indicate the existence of a complementary effect between R&D funding and eco-innovation expenditure.

4.3. Robustness

One of the weaknesses of cross-sectional data models is their limited ability to control for time trends and time-varying characteristics and deal with unobserved individual heterogeneity. In order to address these issues, we used the longitudinal version of the ELE-5 survey to estimate panel data models. This dataset includes a sub-set of firms that were surveyed in 2016 and 2018; that is, they were part of the ELE-4 and ELE-5 surveys. (The ELE-4 survey retrieved firm-level data for the years 2014 and 2015.) Most of the variables used in this empirical research were included in both questionnaires; nevertheless, questions used to proxy eco-innovation decisions were not included in the ELE-4 questionnaire. Given this restriction in the available data, we used firms’ responses to questions C081 and C090 (i.e., current expenditure and total investment on environmental protection in 2016, respectively) from the ELE-5 questionnaire, along with questions C099 and C108, to compute our dependent variables. (Current expenditure and total investment in environmental protection were deflated by the consumer price index in order to express these variables in Chilean pesos in 2017.) In summary, the dependent variables included data from 2016 and 2017, whereas the vectors of covariates considered firm characteristics observed in 2015 and 2017, respectively. Results from a standard random-effects model and fixed-effects with a sample selection model proposed by [52] are shown in Table 6 and

Table 7. Estimates from fixed-effects model with sample selection of current expenditure and total investment in eco-innovation.

Dependent Variable:	Current Expenditure on Eco-Innovation				Total Investment in Eco-Innovation
	Selection Model	Current Expenditure	Selection Model	Current Expenditure	Selection Model	Total Investment	Selection Model	Total Investment
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
CSR policy and dimensions:
CSR policy (intensity index)	0.611 *	−0.004			1.236 *	−0.243
	(0.25)	(0.08)			(0.53)	(0.36)
Environmental CSR			−0.945	0.047			−0.468	0.647
			(0.68)	(0.11)			(0.96)	(0.48)
Financial CSR			−0.541	0.312			1.550	−0.138
			(0.76)	(0.26)			(1.62)	(0.96)
Social CSR			1.663 **	0.027			2.262 *	−0.589
			(0.52)	(0.11)			(1.03)	(0.65)
Firm’s financial performance:
ROA	0.038	−0.189	0.041	−0.177	−0.660	0.011	−0.599	0.088
	(0.19)	(0.17)	(0.18)	(0.16)	(0.45)	(0.37)	(0.52)	(0.42)
Leverage	−0.189	0.335	−0.432	0.322	0.136	0.963	−0.396	0.787
	(1.11)	(0.50)	(1.11)	(0.47)	(2.25)	(1.44)	(2.34)	(1.48)
Business environment:
Importance of competition	−0.093	−0.075	−0.106	−0.072	0.092	−0.281	−0.317	−0.184
	(0.33)	(0.09)	(0.26)	(0.08)	(0.58)	(0.38)	(0.66)	(0.43)
Importance of legal regulations	0.324	−0.043	0.214	−0.021	−0.463	0.090	−0.401	−0.017
	(0.32)	(0.08)	(0.28)	(0.08)	(0.62)	(0.30)	(0.70)	(0.39)
Customer engagement			−0.286	0.150
			(0.53)	(0.19)
Firm characteristics:
Firm size	0.418	0.258	0.306	0.290	2.235	−1.206	2.084	−1.159
	(0.53)	(0.28)	(0.39)	(0.29)	(1.51)	(1.37)	(1.60)	(1.40)
Board of directors’ size	−0.025	0.003	−0.015	0.002	−0.031	−0.004	−0.027	−0.005
	(0.05)	(0.01)	(0.04)	(0.01)	(0.11)	(0.04)	(0.07)	(0.02)
Firm’s export orientation	−0.224	0.149	−0.102	0.093	−0.261	−0.276	0.614	−1.020
	(0.69)	(0.19)	(0.42)	(0.17)	(1.51)	(1.27)	(1.75)	(1.59)
Social media for marketing image	0.902 **	0.281	0.844 *	0.201	−0.645	0.016	−0.290	−0.145
	(0.34)	(0.17)	(0.41)	(0.17)	(1.15)	(0.64)	(1.03)	(0.65)
R&D efforts	−0.290	−0.044	−0.190	−0.036	0.130	0.338	0.152	0.100
	(0.67)	(0.18)	(0.46)	(0.19)	(1.04)	(0.57)	(1.08)	(0.57)
Constant	−2.650 ***	8.681	−2.608 ***	9.969 **	−2.879 ***	7.925	−2.731 ***	7.644
	(0.23)	(5.79)	(0.24)	(3.40)	(0.54)	(5.29)	(0.62)	(4.92)
Observations	8004		8004		8004		8004
No. of firms	4002		4002		4002		4002
No. of bootstrap replications	500		500		170		200

Note: Sectorial dummies are unreported. Standard errors in parentheses were computed using bootstrap replications. * p < 0.05, ** p < 0.01, *** p < 0.001.

, respectively.

Regarding the effect of CSR on eco-innovation decisions among Chilean firms, estimates from random-effects and fixed-effects with sample-selection models confirmed results previously observed; that is, the willingness to allocate resources to eco-innovation is positively influenced by the existence of a general CSR policy and by environmental dimensions of CSR more specifically (columns 1–6, Table 6; columns 1 and 5, Table 7). As previously highlighted, the willingness to spend on and invest in eco-innovation seems to increase with a stronger commitment to transparency and accountability—financial and social dimensions of CSR (column 6, Table 6; columns 3 and 7, Table 7).

In contrast, the business environment, measured through the importance of market competition, emerged as a deterrent to investing in eco-innovation (columns 4–6, Table 6). We have complementary evidence reaffirming that resources allocated to eco-innovation are higher as the size of firms and their boards of directors grow (columns 1–6, Table 6). Finally, investing in eco-innovation is positively influenced by R&D efforts (columns 4–6, Table 6) and a firm’s export orientation (columns 1–6, Table 6). Finally, spending on eco-innovation is larger among firms that use social media for marketing images and products (columns 1 and 3, Table 7).

5. Conclusions

Increasing environmental concerns have encouraged firms to rethink their corporate culture, extending the realm of corporate social responsibility into environmental commitments that can mitigate the negative impact of economic activity on the environment. Research on developed countries has shown a positive effect of environmental CSR on eco-innovation practices among firms; however, less evidence has been reported for developing countries and for different dimensions of CSR. Based on the triple bottom line perspective, this study used firm-level data to address the relationship between environmental, financial, and social dimensions of corporate culture and the willingness to spend on and invest in eco-innovation among Chilean firms. The following remarks can be highlighted from our findings.

The evidence from binary probit, Tobit, and Heckman sample-selection models confirmed a positive effect of CSR policy on the willingness to eco-innovate and on the resulting resource allocation decision. In turn, analyzing the multidimensional character of corporate culture revealed that when firms report a greater commitment to environmental and financial CSR, their willingness to eco-innovate is higher. That said, this increased commitment does not affect the amount of resources the firm devotes to eco-innovation practices. Current expenditure on eco-innovation only increases when the social dimension of CSR exhibits a stronger commitment to transparency among stakeholders and the community. That said, this characteristic of social CSR seems to reduce the willingness to eco-innovate in the first place. This result may signal that social CSR is complementary for boosting both expenditure and investment in eco-innovation once the decision to eco-innovate is already made. Together, these results indicate that among eco-innovators, social CSR may generate synergies with other CSR dimensions. Thus, we shed some light on the existence of a complementary effect between CSR dimensions in developing contexts, a less-explored research question that has been previously remarked upon in the related literature [18].

The effect of financial performance on eco-innovation is conditional upon the indicator used to analyze the relationship. While expenditure on eco-innovation is positively driven by the ROA ratio, the total investment in environmental innovation is negatively influenced by financial debt indicators, i.e., leverage ratios.

Our estimates also revealed that the relevance of legal regulations in firms’ growth perspectives, firms’ export orientation, and firms’ R&D efforts are positive drivers of eco-innovation among Chilean firms. Larger-sized companies exhibit a higher disposition to eco-innovate, allocating larger amounts of expenditure and investment in environmental innovation. Therefore, a pro-environmental policy should focus on micro- and small-sized companies, which confront higher existing entry barriers in the process of adopting environmentally friendly practices.

Finally, corporate governance practices seemed to influence the inclination to eco-innovate among Chilean firms. The willingness to eco-innovate is positively influenced by the board of directors’ size, but, at the same time, a larger board size is also a detriment to how much gets spent on eco-innovation among Chilean firms.

6. Limitations and Further Research

The following limitations can be highlighted from our empirical study. The questions used to proxy eco-innovation considered an input perspective, with emphasis on innovation in pollution prevention, omitting aspects linked to other eco-innovation indicators in the form of immediate output, direct output, indirect impact, and firm performance. Note that the Chilean Survey of Enterprise Innovation may allow for addressing these issues; nevertheless, questions related to the implementation of CSR policies among Chilean firms are not included in the ENI survey. Therefore, the scope of the definition of eco-innovation was narrowed due to data availability regarding our variable of interest. Future research could extend this analysis by considering these alternative approaches for measuring eco-innovation.

Moreover, the adoption of CSR policies could eventually be endogenous to the willingness to eco-innovate, which would imply extending the empirical strategy. In this regard, future research could explore the use of the conditional mixed-process framework, which was proposed by [53] for a broader range of seemingly unrelated regressions (SUR) models.