Mitigating Involutionary Competition Through Corporate ESG Adoption: Evidence from the Consumer Electronics Manufacturing Industry

Abstract

This study investigates whether and how corporate commitment to environmental, social and governance (ESG) performance can mitigate involutionary competition in China’s consumer electronics manufacturing industry. By constructing a quantifiable index of involutionary competition intensity and matching it with corporation-level ESG scores, we document a statistically significant negative association between ESG performance and the degree of involutionary competition. Mechanism analysis reveals that ESG mitigates involutionary competition through two primary channels: (1) differentiation strategies that reduce price-based competition and product homogeneity, and (2) market-order regulation that curbs opportunistic behaviour and raises R&D efficiency. A modest price increase is shown to be revenue-enhancing; moreover, random-forest simulations indicate that counter-involutionary competition efforts amplify the market-share gains from cooperative R&D expenditures, accelerating post-adjustment revenue growth. This transition generates simultaneous increases in corporate profits and corporation value, breaking the previous price ceiling and establishing a sustainable development loop. The findings provide actionable insights for shifting the industry from low-level rivalry to sustainable value creation.

1. Introduction

In recent years, the consumer electronics manufacturing industry in China has witnessed an increasingly severe phenomenon of involutionary competition—a self-reinforcing cycle of low-price, low-margin rivalry characterized by product homogenization, capacity over-expansion, and diminishing innovation returns. Rooted in structural oversupply, convergent technological trajectories, and sub-national “race-to-the-bottom” investment incentives, this dynamic has eroded profitability, distorted resource allocation, and trapped corporations in a zero-sum game [1]. While prior literature has extensively documented the symptoms of involutionary competition—such as price wars, compressed margins, and duplicative R&D—systematic investigations into viable mitigation strategies remain scarce. This paper addresses this gap by exploring whether and how corporate commitment to environmental, social, and governance (ESG) performance can serve as a strategic lever to counteract involutionary pressures.

The theoretical rationale for an ESG-based intervention rests on two interrelated arguments. First, ESG initiatives compel corporations to internalize externalities, shifting managerial focus from short-term cost competition to long-term value creation [2]. Second, ESG-driven differentiation strategies (e.g., green innovation, supply chain transparency, and stakeholder governance) create non-replicable competitive advantages that dilute price-based rivalry [3]. However, empirical validation of these mechanisms within the context of involutionary competition is limited, particularly in the consumer electronics sector, where rapid product cycles and intense cost pressures exacerbate the tension between sustainability commitments and competitive survival.

Practically, this study delivers a usable toolkit for firms and regulators to shift from price wars to durable, capability-driven growth. It offers a replicable product-level benchmarking routine using launch prices and similar performance groupings to quantify profit erosion, calibrate brand premiums, and support portfolio decisions, filling an operational gap noted in market-design discussions. It further operationalizes pricing and R&D choices with ARIMAX demand models for revenue forecasting under moderate price increases, GAM based breakpoint detection to locate elasticity thresholds such as the 8000 yuan regime, and random forest simulations that translate R&D outlays with a 1.5 billion yuan attenuation threshold into market share gains and multi year revenue trajectories, which complements empirical practice in adjacent product and innovation analytics [4,5]. Finally, it informs policy by specifying disclosure and due diligence standards and by motivating shared carbon footprint computation platforms that reduce verification frictions and shift competition toward auditable performance, aligning with regulatory aims highlighted in [6,7].

2. Literature Review

2.1. Research on ESG Strategy for Corporations

Contemporary corporate ESG strategy research exhibits characteristics of multidimensional integration and convergence. Scholars focus on evaluating the effectiveness of ESG implementation, such as conducting a systematic analysis of commercial banks’ ESG practices using the Performance Prism Model [8]. Research has found that smart manufacturing significantly enhances corporate ESG performance through technological empowerment [9], while industrial internet strengthens the competitiveness of small and medium-sized corporations through a technology-organization-environment synergy mechanism [10]. In the field of climate action, research has revealed that corporations need to formulate interdependent sustainable decisions [11], asset owners are dynamically adjusting their climate investment strategies to address transition risks [12], and foreign investors’ climate risk exposure assessments have become a key strategic consideration [13]. In terms of innovative technology applications, circular economy models such as biochar-catalyzed conversion of plastic waste have been proven to drive sustainable transformation [14]. These studies collectively indicate that an effective ESG strategy must integrate technological empowerment, cross-value chain collaboration, precise quantitative assessment, and climate risk management to form a multi-dimensional, interconnected systemic implementation framework.

2.2. Research on Corporate Price Competition

Contemporary research on corporate price competition exhibits multi-dimensional interconnections, with non-price factors profoundly reshaping the competitive landscape: studies have found that employee stock ownership plans indirectly influence pricing strategies by optimizing corporate governance [15], while ESG ratings have become a core pathway for enhancing competitiveness by driving digital transformation and green innovation in logistics corporations. Digital finance, meanwhile, can effectively curb ESG decoupling to strengthen substantive competition [16]. At the supply chain level, the role of midstream entities in promoting sustainable practices [17] highlights the value of supply chain collaboration, while late-entering corporations achieve catch-up through standardized technology strategies [18]. Research on market mechanism design warns that power sector reforms must balance efficiency with regulatory capture risks, while public attention to the nonlinear regulatory effects of ESG and green innovation and energy entrepreneurship vitality [19] collectively reveal the complex interactive mechanisms between policy environments, social pressures, and entrepreneurial spirit in constructing a sustainable competitive ecosystem.

2.3. Research on the Correlation Between ESG and Corporate Competitiveness

Research on the correlation between ESG and corporate competitiveness reveals a digitally driven deep integration mechanism: digital transformation empowers manufacturing corporations to transform into service-oriented businesses and enhances their innovation capabilities [20], effectively curbing ESG decoupling [21] and improving the ESG performance of small and medium-sized corporations [22]. Research has found that ESG practices enhance customer stability through external regulation and reputation mechanisms [23], but there is a potential risk of disconnect between corporate environmental communication and green innovation actions [24], while digitalization can moderate the constraints imposed by environmental tax laws on greenwashing. Green financial tools such as green credit significantly improve the environmental performance of high-polluting corporations [25], highlighting the synergistic effects of policy regulation and market incentives. These findings collectively indicate that ESG competitiveness stems from digital capability building, innovation ecosystem construction, and substantive practices driven by both regulation and market forces. It systematically reconfigures corporate sustainable competitive advantages by reducing compliance costs, optimizing stakeholder relationships, and mitigating greenwashing risks.

2.4. Literature Summary and Commend

Existing research shows that ESG enhances firm competitiveness via digital transformation, green innovation, financing advantages, and governance improvements, and that non price capabilities can reshape competition across supply chains and sectors. However, while prior research has linked ESG to competitiveness and innovation, its specific role in breaking cycles of involutionary competition, characterized by profit erosion from homogeneous price wars, remains unexplored. In particular, the literature under theorizes and under measures two core channels for escaping involution: ESG enabled product differentiation (e.g., verifiable low carbon design, circular models, privacy by design) that sustains price premia, and market order governance (e.g., supply chain due diligence, anti corruption transparency, standardized disclosures) that raises the cost of opportunism and curbs predatory pricing. This paper addresses that gap by constructing a product level involution intensity index, matching it with firm level ESG metrics, and identifying a robust negative ESG involution relationship, then tracing the effect through the differentiation and market order mechanisms using econometric, time series, and machine learning evidence.

3. Measurement of Involutionary Competition and Its Relation with ESG

Before conducting research on the relationship between corporations’ ESG pursuits and involutionary competition, it is necessary to transform the abstract concept of “involutionary competition” into a concrete, quantifiable “degree.” The concept of involutionary competition can be broken down into factors such as “intense price competition, low profit margins, high inventory turnover, high cash discount rates, extended accounts receivable collection periods, severe product homogenization, and low conversion rates of R&D expenditures into innovative outcomes.” Among these, price competition is the most direct and significant indicator of involutionary competition. Therefore, this study will measure involutionary competition from the perspective of profit erosion caused by corporations proactively lowering prices.

In the study, the representative business of mobile phone sales in the consumer electronics manufacturing industry was selected, and TOB-end supplier corporations were excluded to control the impact of average profit margins. We also selected Apple smartphones (Cupertino, CA, USA), which currently have the highest average premium under the same performance, as a benchmark. We compared products with similar performance parameters (ref. [26] for the application of such technique) to Apple and defined this group of smartphones with similar performance as the “similar performance parameter price comparison group.” Within this group, the smartphone with the highest price excluding Apple has very low involutionary competition, and the price difference between it and Apple smartphones is approximately equal to Apple’s brand premium. The price difference between the smartphone with the lowest involutionary competition and any other smartphone represents the profit portion that other smartphones voluntarily sacrifice to gain market share through price competition. (3) Additionally, since prices have regional variability, fluctuate with new product iterations, and are influenced by macroeconomic factors such as current national policy subsidies in China, which can lead to actual consumer discount rates and changes in actual payments, the product launch price is selected as the base research metric to minimize the impact of macroeconomic market fluctuations on prices. (4) Finally, we decided to quantify the degree of involutionary competition using a relative metric rather than an absolute value to mitigate the effects of differences in corporation size and the intensity of competition across high-, mid-, and low-end segments. As shown in Formula (1), this approach enables the construction of an evaluation model for measuring the degree of involutionary competition.

$$\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{i}\mathrm{t} \mathrm{d}\mathrm{e}\mathrm{c}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e} \mathrm{d}\mathrm{u}\mathrm{e} \mathrm{t}\mathrm{o} \mathrm{i}\mathrm{n}\mathrm{v}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{a}\mathrm{r}\mathrm{y} \mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{e}\mathrm{t}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}={\displaystyle \frac{P_{MAX}^{GROUP}-P_{\propto }}{P_{iPhone}}}$$

(1)

Based on the corporations selected for mobile phone operations, 13 representative corporations in China’s consumer electronics manufacturing industry were selected. Two ST* corporations were excluded to reduce the abnormality of the research data. The sales data of their main product categories from 2020 to 2024 was compared with that of Apple’s iPhones with similar performance parameters to infer the degree of involutionary competition.

3.1. Construction of a Comparative Scoring System

The reduction in corporate business profits caused by involutionary competition cannot be generalized, as it is also influenced by the profit level characteristics of different product types within the same business. Therefore, when comparing involutionary competition, it is necessary to categorize mobile phones with similar performance parameters into comparison groups and then compare their price levels within the group to accurately calculate the profits they have voluntarily abandoned due to involutionary competition.

3.1.1. Product Performance Parameter Evaluation

When constructing the parameter scoring system (refer to [27,28] for similar application of parameter scoring system), we extensively referenced the professional methodologies of multiple authoritative evaluation institutions to ensure the system’s scientific rigor and comprehensiveness. This parameter scoring system draws on the objective-subjective combination approach of DxOMark’s imaging and display evaluations, DisplayMate’s multi-dimensional display testing standards, AnandTech’s hardware performance evaluation methodology, and Tom’s Guide’s multi-dimensional comprehensive assessment model, among other authoritative methodologies. It covers dimensions such as hardware performance, imaging systems, display quality, battery life and charging, storage configuration, user experience (operating system, audio, and thermal management), and ecosystem value (ecosystem integration and after-sales service). After collecting the specific parameters of each smartphone, we evaluate its performance parameters based on specific score weights.

The specific weighting distribution is divided into six parts: (1) Hardware performance dimension, as the core foundation of product operation, directly impacts critical experiences such as multitasking and running large applications. We use AnandTech’s authoritative benchmarking data to quantify processor performance and test memory and GPU performance, with a weighting of 20 points. (2) Imaging System Dimension: With the widespread adoption of mobile photography, camera resolution has become a key consideration for users when selecting a device. The main camera score is referenced from DXOMARK’s imaging rankings, while the secondary camera and video recording are scored based on functional completeness and calibration quality, with a weight of 18 points. (3) Display Quality Dimension: This dimension is further broken down into resolution, refresh rate, brightness, and color accuracy. The evaluation references DisplayMate’s testing standards and evaluation methods for each metric, Weighting: 15 points. (4) Battery life and charging dimension: Scored based on industry-standard benchmarks and technical capabilities, with a weighting of 10 points. (5) Storage configuration dimension: Scored based on industry-standard benchmarks and technical capabilities, with a weighting of 10 points. (6) User experience dimension: Operating system smoothness and feature richness are evaluated based on actual user feedback and industry analysis. Industrial design is objectively evaluated based on material quality and craftsmanship, weight, and thickness. Audio performance and cooling systems are referenced from professional reviews and actual test results. Weight: 15 points. (7) Ecosystem Value Dimension: Ecosystem synergy is assessed based on multi-device connectivity capabilities. After-sales service is evaluated based on global warranty coverage and the density of offline service centers. Weight: 12 points. Specific scores for each smartphone model are listed in Appendix A.

3.1.2. Derivation of Corporate ESG Scores

This study adopts the ESG scores provided by the Wind database as the primary metric because the underlying methodology is event-driven, empirically grounded and continuously self-updating. Corporations lacking publicly disclosed ESG data were assigned scores by replicating Wind’s “data-driven, locally adapted and dynamically penetrative” framework.

Table 1. Huawei ESG Score Calculation for the Second Quarter of 2020.

Calculation Items	Key Events	Assessment Stratification	Score Breakdown	Project Proportion	Project Score
E	Greenhouse Gas Emissions	Data Performance	Total carbon emissions amounted to 85,000 metric tons of carbon dioxide equivalent (t CO2e), with a carbon emissions intensity of 3.2 t CO2e per million dollars of revenue. Compared to the previous quarter, this represents a decrease of 4.5%.	30%	6.5
		Industry Comparison	Compared to industry peers such as Samsung Electronics and ZTE Corporation, the carbon emissions intensity remains at a relatively high level; Samsung Electronics’ carbon emissions intensity for the quarter was 2.8 t CO2e per million USD in revenue, while ZTE Corporation’s was 2.5 t CO2e per million USD in revenue.
		Goal Achievement	The corporation plans to reduce carbon emissions per unit of revenue by 40% over the next 10 years.
	Energy Management	Venue Energy-Saving Measures	Over 70% of the lighting systems in the headquarters office area have been upgraded, resulting in a 22% reduction in lighting energy consumption compared to the previous quarter. Additionally, intelligent temperature control technology has been introduced in the air conditioning systems, reducing energy consumption in office spaces by 15%.	40%	7.5
		Data Center Energy Optimization	Self-developed intelligent energy management systems have been deployed in multiple large-scale data centers, with energy utilization rates in these data centers increasing by 8% compared to the previous quarter, and energy consumption per unit of computing power decreasing by 6%.
		Renewable Energy Utilization	In a data center in a European country, a long-term power purchase agreement was signed with a local renewable energy supplier, increasing the proportion of renewable energy used from 30% to 45% in this quarter. In China, distributed solar photovoltaic power generation equipment has also been installed in some office buildings and data centers, with photovoltaic power generation accounting for 12% of total electricity consumption in this quarter.
	Resource Recycling and Utilization	Green Product Design	The new smartphone series features a modular design that facilitates disassembly and component recycling at the end of the product life cycle. In terms of raw materials, 35% of the phone casings are made from recycled plastic. Material substitution technology has been used to reduce resource consumption during the production process. According to an assessment, the green design ratio of the new smartphone series has increased by 15% compared to the previous generation.	30%	7
		Resource Optimization in Production Process	In the manufacturing process, Huawei has optimized production processes and supply chain management to reduce raw material waste; introduced advanced production management systems to achieve fine-grained control; and worked closely with suppliers to promote optimization of raw material packaging and reduce the use of packaging materials. This quarter, the amount of raw materials used per unit of product in the manufacturing process decreased by 5% compared to the previous quarter.
		Recycling System Construction	Continuously improved the product recycling system: In China, established in-depth partnerships with a number of professional recycling corporations to expand recycling channels; online, set up convenient recycling access points on Huawei’s official website and mobile app; offline, set up recycling points in Huawei stores in major cities; overseas, cooperated with well-known local recycling organizations to carry out electronic product recycling, with a waste product recycling rate of 28%, an increase of 3 percentage points over the previous quarter.
S	Protection of Employees’ Rights and Interests	Training and Development	We have enriched and improved the employee training system, with the average training time per employee reaching 28 h, an increase of 5 h compared to the previous quarter. We have provided customized leadership training courses for employees with promotion potential.	40%	8
		Compensation, Benefits and Satisfaction	The corporation’s compensation levels are competitive within the industry, with the addition of supplementary commercial insurance and flexible working arrangements. Employee satisfaction survey results show that employee satisfaction has increased by 3% compared to the previous quarter, reaching 85%.
		Health and Safety Assurance	Installed advanced fire alarm systems and emergency evacuation equipment; strengthened employee safety training, organized regular safety drills to enhance employees’ safety awareness and emergency response capabilities; provided health consultation and psychological counseling services to employees; and established health stations in some office locations.
	Supply Chain Social Responsibility	Supplier Audit	Social responsibility audits were conducted on 80% of first-tier suppliers, covering areas such as labor rights, environmental protection, and workplace safety. Audit results revealed that 8 suppliers had instances of overtime work, and 4 suppliers had insufficient investment in environmental protection facilities.	30%	6.5
		Rectification and Assistance	For suppliers with overtime issues, Huawei collaborated with them to analyze production processes and personnel allocation, and reasonably scheduled employee working hours. For suppliers with insufficient environmental protection facility investments, Huawei organized expert teams to provide environmental protection technical solutions and assist suppliers in upgrading and renovating their equipment.
	Community Participation and Public Welfare Activities	Educational Public Welfare Projects	Active educational welfare programs globally. In a certain African country, Huawei collaborated with local governments and educational institutions to launch a digital education empowerment program (the project donated communication equipment, smart teaching terminals, and online education platform services to 20 schools), benefiting over 5000 students	30%	7
		Environmental Public Welfare Activities	Large-scale environmental public welfare activities were organized. With the theme of “Green Action, Guarding Our Home”, the total number of trees planted in this quarter reached 15,000. At the same time, garbage classification publicity activities were carried out through a combination of online and offline methods, covering 80 communities, and the number of directly benefited residents exceeded 30,000. These activities have enhanced the public’s environmental awareness and promoted the improvement of the local ecological environment.
G	Corporate Governance Structure	Operation of Governance Entities	All governance entities in the corporate governance structure operate in a standardized and efficient manner. In this quarter, the board of directors conducted in-depth discussions on major issues such as the corporation’s business development direction and market expansion strategies; the board of supervisors earnestly performed its supervisory duties to ensure that the corporation’s operations comply with laws, regulations and the articles of association.	30%	8
		Transparency of Decision-Making Process	The decision-making process is highly transparent. When formulating a new product research and development strategy, opinions are widely collected through internal forums, expert consultation meetings, etc., and after multiple rounds of demonstration and evaluation, it is submitted to the board of directors for deliberation. The corporation promptly discloses decision results and related information to internal employees and external stakeholders.
	Risk Management and Compliance	Risk Identification and Assessment	Through the risk early warning mechanism and regular risk assessment meetings, various risks including market risks, technical risks, compliance risks, etc., have been identified; the risk of declining market competitiveness caused by lagging technological innovation has been identified; the compliance risks brought by differences in laws and regulations in different countries and regions have been identified.	40%	7.5
		Risk Response Measures	In this quarter, through effective risk response measures, some potential risks were successfully resolved, and the impact of policy adjustments in a certain country on the corporation’s business was dealt with in advance, avoiding major losses.
		Effectiveness of Compliance Management	Compliance training was organized for all employees, with a training coverage rate of 100%. The corporation actively cooperated with the inspections and audits of government departments in various countries, and no major compliance issues occurred.
	Information Disclosure and Transparency	ESG Report Quality	The content of the report is comprehensive and detailed, covering goals, strategies, practices and achievement data in various aspects such as environment, society and governance.	30%	8
		Other Information Disclosure	The corporation’s major information is promptly disclosed through various channels such as the official website, press conferences, and investor relations platforms.
Overall score	$30\%\mathrm{E}+35\%\mathrm{S}+35\%\mathrm{G}$				7.46

outlines the 2022-Q2 ESG derivation for Huawei (Shenzhen & Dongguan, China) as an illustrative case. Also based on this scoring methodology, the ESG scores of other unlisted corporations for which no data are available are estimated, as shown in Figure 1.

3.2. Measuring the Degree of Involutionary Competition Based on Profit Erosion

Calculations based on Formula (1) involve grouping products with similar scores and a score difference within 7 percentage points, while excluding samples where the score difference exceeds 7 percentage points or the release time difference exceeds 12 months. The ≤7-percentage-point score gap and ≤12-month release window align with de facto handset industry comparability standards: (1) Performance parity: cross-benchmark dispersion for same-generation flagships (Geekbench/AnTuTu/GFXBench/PCMark) typically spans 2–5 pp due to firmware, thermal headroom, and memory bins; setting 7 pp clears measurement noise yet stays below the 8–15 pp jump that OEMs use to separate adjacent tiers (e.g., Snapdragon 8 vs. 8+; Dimensity “+” refreshes), preserving a single-tier peer set. (2) Carrier/OEM ranging: operators and retail planograms define competitive sets by model-year; flagship and upper-mid cycles are annual, with subsidies, warranties, and OS roadmaps keyed to a 12-month cadence, so a ≤12-month window reflects real shelf competition while avoiding cost-curve breaks from new silicon nodes. (3) Pricing integrity: launch MSRP is the industry reference for positioning during the first year; beyond 12 months, lifecycle discounting and channel promotions dominate, confounding profit-erosion attribution. (4) Supply-side coherence: major SoC, camera sensor, and modem generations update roughly yearly; extending cohorts past 12 months mixes different BOM/yield regimes. Sensitivity checks (5/10 pp; 9/15 months) are recommended to confirm robustness while staying within these standards.

Prices within each group are then compared pairwise, and the differences are calculated. These differences are further divided by the corresponding Apple iPhone prices to determine Apple’s brand premium and the profit decline rate of the respective smartphone manufacturers due to involutionary competition. Specific results are shown in

Table 2. Mobile Phone Internal Turnover Measurement Scale.

Group Number	Year Quarter	Apple Model	Comparable Model	Profit Decline Rate due to Involutionary Competition
1	2020 Q4	iPhone 12 Pro	Huawei P40 Pro	0.00%
			Xiaomi 10 Pro	11.64%
			Meizu 17 Pro	19.87%
			Nubia Z20	29.29%
2	2020 Q4	iPhone 12	iQOO 3	0.00%
			realme X50 Pro	6.33%
			OPPO Reno4 Pro	3.16%
3	2021 Q4	iPhone 13 Pro	vivo X70 Pro+	0.00%
			OnePlus 9 Pro	6.25%
			Honor Magic3	11.25%
			Motorola edge s pro	38.75%
4	2021 Q4	iPhone 13	Xiaomi 12 Pro	0.00%
			realme GT	23.34%
			Nubia Z30 Pro	21.67%
5	2022 Q4	iPhone 14 Pro	Xiaomi 12S Ultra	0.00%
			iQOO 9 Pro (Vivo Mobile Com-munications Co., Ltd., Dongguan, China)	12.50%
			Honor Magic4 Pro	6.25%
			Huawei Mate 50 Pro	6.25%
6	2022 Q4	iPhone 14	realme GT2 Pro	0.00%
			Nubia Z50	15.00%
7	2023 Q4	iPhone 15 Pro	vivo X90 Pro+	0.00%
			Xiaomi 13 Ultra	6.25%
			ZTE Nubia Z60 Ultra	31.25%
8	2023 Q4	iPhone 15	Huawei P60 Pro	0.00%
			OnePlus 11	49.82%
			realme GT3	61.49%
9	2024 Q4	iPhone 16 Pro	vivo X200 Pro	0.00%
			OPPO Find X7 Ultra	5.00%
			Honor Magic7 Pro	12.00%
10	2024 Q4	iPhone 16	Xiaomi 14 Ultra	0%
			Transsion TECNO Phantom X3	43.07%

.

3.3. Empirical Analysis

(1) Correlation Analysis

The degree of involutionary competition and the corresponding ESG scores for each quarter have been obtained. The next step is to verify whether there is a correlation between the two. To test the baseline hypotheses of this study, we use the two-way fixed effects model (A two-way fixed effects model is appropriate because it controls for unobserved, time-invariant firm heterogeneity (firm FE) and common time shocks (time FE), which are likely to confound ESG and outcomes in a small, heterogeneous, and possibly unbalanced panel. Identification comes from within-firm variation over time, a more credible source than cross-sectional differences with n ≈ 31. TWFE remains consistent under unbalancedness if missingness is ignorable, and with firm-clustered SEs (using small-sample adjustments or wild-cluster bootstrap) it yields robust inference despite short T) to identify the impact of ESG ratings on the within-corporation pay gap [29]. The basic model is as follows:

$${\mathrm{y}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{x}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{x}1}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{x}2}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4{\mathrm{x}3}_{\mathrm{i}\mathrm{t}}$$

where y_it is the value of the i-th internal interest rate at time t, x_it is the ESG score, x1_it is the years since establishment, x2_it is the Industry Lerner Index for the i-th variable at time t, and x3_it is corporation size at time t for the i-th variable (i is the i-th variable and t is the variable at time t), as shown in

Table 3. Notations in the model equation.

Notation	yit	xit	x1it	x2it	x3it
Meaning	Value of internal interest rate	ESG score	years since establishment	Industry Lerner Index	corporation size

below. Control variables selected include: years since establishment, industry Lerner index, and corporation size.

Table 4. Variable correlation analysis.

	y	x	x1	x2	x3
y	1
x	−0.445 **	1
x1	−0.087	0.226	1
x2	−0.068	−0.465 ***	−0.11	1
x3	−0.508 ***	0.566 ***	0.400 **	−0.169	1

** p < 0.05, *** p < 0.01.

presents pairwise correlations among the variables, where its computation follows a flowchart shown by Figure 2 below. The Pearson coefficient between the explanatory variable—corporate ESG score—and the dependent variable (involutionary competition concession ratio) is −0.445 and is statistically significant at the 5% level, providing prima facie evidence of a negative association between ESG performance and involutionary competition intensity, where we referred to [30,31,32] for similar analysis techniques. Corporation size exhibits a −0.508 correlation with the involutionary competition ratio (significant at 1%) and a +0.566 correlation with the ESG score (also 1%), while its correlation with corporation age is +0.400 (5%). The industry Lerner index is negatively correlated with the ESG score (−0.465, 1%). The absolute magnitudes of most coefficients lie around 0.5, alleviating concerns of severe multicollinearity; nevertheless, the adequacy of a linear specification remains an open question, prompting the baseline regressions that follow.

(2) Benchmark Regression Analysis

Table 5. Benchmark regression.

	(1)	(2)	(3)	(4)
Variable	y	y	y	y
x	−0.059 ***	−0.060 ***	−0.082 ***	−0.053 **
	(0.012)	(0.013)	(0.020)	(0.021)
x1		0.000	0.000	0.002
		(0.003)	(0.003)	(0.003)
x2			−6.749 *	−5.845
			(3.836)	(3.489)
x3				−0.688 *
				(0.338)
_cons	0.546 ***	0.545 ***	1.542 **	1.272 **
	(0.100)	(0.102)	(0.598)	(0.556)
N	31	31	31	31
R2	0.198	0.199	0.295	0.382

Note: The values in parentheses are robust standard errors. ***, **, and * indicate significance levels of 1%, 5%, and 10%, respectively.

reports the baseline ordinary-least-squares results for the effect of ESG scores on the involutionary competition concession ratio. Column (1) presents the parsimonious specification with the core regressor alone; the ESG coefficient is −0.059 and significant at the 1% level. Columns (2) and (4) sequentially incorporate control variables (corporation age, Lerner index and corporation size). Across all specifications the ESG coefficient remains negative and statistically significant at the 5% level or better. The result indicates that higher ESG scores attenuate product-level involutionary competition. Two complementary mechanisms underlie this relationship: first, ESG initiatives divert managerial attention and financial resources toward environmental governance, social obligations and governance improvements, thereby constraining the slack available for homogeneity-inducing price competition; second, corporations prioritising ESG performance are predisposed to pursue technological innovation and product differentiation, substituting sustainable competitive advantages for predatory pricing strategies. The coefficient on corporation size is consistently negative and significant, reflecting both economies of scale that reduce marginal costs and risk-averse managerial preferences among large incumbents for non-price instruments such as brand premiums and channel control.

(3) Robustness analysis

To assess robustness, we conduct a sample-sensitivity test by restricting the estimation window to 2020–2023, thereby excluding potentially influential 2024 observations. Within this truncated sample the ESG coefficient retains its negative sign and remains significant at the 5% level, while its magnitude closely mirrors the baseline estimate. See

Table 6. Robustness test.

	(1)
Variable	y
x	−0.045 **
	(0.021)
x1	0.001
	(0.003)
x2	−5.723
	(4.935)
x3	−0.635 *
	(0.341)
_cons	1.210 *
	(0.692)
N	26
R2	0.343

Note: The values in parentheses are robust standard errors. **, and * indicate significance levels of 5% and 10%, respectively.

below. All other coefficients remain qualitatively unchanged, corroborating the conclusion that corporate ESG performance mitigates product involutionary competition.

4. Feasibility Analysis of ESG Engagement as a Mitigation Pathway for Involutionary Competition

Driven by demand–supply imbalances, convergent technological trajectories and sub-national “race-to-the-bottom” investment incentives, corporations in the consumer electronics sector exhibit involutionary behaviour characterised by low-price homogeneous competition, capacity over-expansion and sub-optimal R&D allocation. ESG disclosure compels corporations to internalise externalities, triggering managerial reforms that redirect resources toward environmental performance, social responsibility and governance compliance in pursuit of higher ESG ratings. These efforts attenuate involutionary competition through three mutually reinforcing channels: (1) differentiation strategies that dilute price-based rivalry, (2) market-order regulation that suppresses opportunistic conduct, and (3) By summarizing and analyzing the major governance events of consumer electronics manufacturing corporations and the response strategies of each corporation, as shown in

Table 7. ESG summarises the ESG–involutionary competition mitigation pathways across the three dimensions.

Pursuit Category Actions Conditions Affecting Internal Roll-Up		Increase R&D Expenses	Decrease Inventory Turnover Rate	Improve Unit Profit Margin	Increase Product Prices	Rising Labor Costs	Increased Capital Expenditures	Taxes Reduced on a Project-by-Project Basis
E	Green material research and development
	Carbon-neutral factory certification
	Low-carbon construction initiatives
	Purchase of environmentally friendly equipment
	Water recycling systems
S	Employee salary increases
	Barrier-free production line construction
	Enhanced privacy protection
	Rural education partnerships
G	Rural education partnerships Blockchain-based anti-corruption systems
	Circular business models
	ESG data platforms
	Open-source technology patents

, the path of action by which a corporation’s pursuit of ESG can mitigate involutionary competition can be explored.

4.1. Alleviating Price Competition

4.1.1. Environmental Dimension (E)

Environmental stewardship in consumer electronics manufacturing can convert compliance expenditures into pricing-power advantages, thereby mitigating low-price involutionary competition. Instruments include green-material R&D, zero-carbon factory certification, low-carbon supply chain retrofitting, advanced abatement equipment and closed-loop water systems.

(1) Green materials

Corporations that industrialise green materials at scale dilute unit costs and expand margins, supporting terminal-product premiums. Apple-supplier Shandong Innovation Group’s (Jinan, China) recycled-aluminium technology achieves 99.95% grade-purity regeneration and a 42% energy-intensity reduction, lowering Mac enclosure material costs by 22% and lifting Apple’s unit margin to 59%.

(2) Zero-carbon certification

Mandatory standardised audits compress environmental-review cycles and accelerate inventory turnover, releasing working capital that offsets initial outlays. Certification also unlocks state subsidies and green-premium purchase orders, expanding unit margins by up to 9 pp and conferring autonomous pricing power that obviates price wars. Reinvested cash-flows finance on-site PV integration and energy-efficiency retrofits, entrenching a virtuous “invest–save–premium” cycle.

(3) Low-carbon supply chains

Certified low-carbon supply chains raise inventory-turnover ratios and cut customs-clearance time by 40%, with warehousing-cost savings covering 85% of certification expenses. Carbon-labelling premiums and CBAM exemptions jointly expand unit margins by 6.2 pp. Capital released from efficiency gains is redeployed into carbon-data platforms (ref. [33]) that further compress logistics and maintenance cycles, driving price elasticity below the sector median.

(4) Abatement equipment

Upgraded abatement equipment reduces specific energy consumption, differentiating cost structures and weakening the need for cost-pass-through. Concurrent tax credits and green subsidies raise net-present-value rates to 23.5% and shorten pay-back to 2.5 years, obviating capacity expansion for cost dilution.

(5) Water-recycling systems

Water-recycling investments qualify for R&D subsidies and a 50% water-resource tax rebate, raising unit margins by 7.1 pp. Recycled-water savings are reinvested, triggering further cost reductions and extending premium cycles beyond price-war horizons.

4.1.2. Social Dimension (S)

In the consumer electronics manufacturing sector, the corporation of humanistic care initiatives within the Environmental, Social, and Governance (ESG) framework—particularly under the “Social” pillar—can be transmuted into sustainable competitive advantage through the dynamic coupling of organizational resilience and dynamic capabilities, thereby disrupting the involutionary trap induced by technological paradigm convergence and product homogeneity.

(1) Wage Premium and Human Capital Appreciation

Elevating compensation levels in alignment with pay-equity principles and long-term incentive (LTI) schemes constitutes a salient social-dimension intervention. This practice enhances employee stickiness, compresses turnover-related transaction costs, and reduces training amortization, leading to an appreciation of corporation-specific human capital. The resultant increase in process yield and unit-level profitability yields a long-term margin expansion of approximately three percentage points, while the emergent human-capital premium secures non-price-sensitive market recognition.

(2) Inclusive Production Lines and Barrier-Free Manufacturing

The deployment of barrier-free production lines operationalizes the normative constructs of inclusive development and equitable employment. By integrating universal-design principles into process architecture, corporations unlock the latent potential of a diversified talent pool, thereby attenuating homogeneity in technological trajectories. This heterogeneity-induced recombinant innovation enhances R&D intensity and accelerates technological iteration. Concurrently, corporations accrue policy-derived economic rents: procurement preferences enlarge order premia, disability-employment levy exemptions reduce labor costs, and earmarked subsidies offset retrofitting expenditures—collectively mitigating price-based competition and elevating industry-wide profit margins.

(3) Privacy-Preserving R&D and Data-Sovereignty Differentiation

Augmenting investment in privacy-preserving technologies functions as a dual-path mechanism—simultaneously erecting technological moats and attenuating regulatory-compliance risk. By minimizing the probability and severity of data-breach contingencies, corporations evade the low-price homogeneity trap and cultivate a reputation for reliability that attracts quality-sensitive clientele. This risk-adjusted strategic posture enables scientifically calibrated capacity planning and averts disorderly expansion, thereby reconciling security expenditure with commercial return and advancing both sectoral profitability and sustainable development capacity.

(4) Rural Skills-Upskilling and Reverse Industrial Subsidization

Implementing rural skills-upskilling programs operationalizes the ESG social mandate of common prosperity and industrial feedback loops. Customized vocational training activates latent rural human resources, optimizes the corporation’s labor-mix structure, and alleviates talent-supply bottlenecks. Co-creating curricula with tertiary institutions fosters industry-education integration, compresses skill-iteration cycles, and raises R&D intensity—counteracting technological convergence and propelling differentiated competition. Fiscal instruments—tax incentives and localized training subsidies—reduce program cost pressure, facilitating human-capital appreciation, margin accretion, and the transcendence of involutionary dynamics.

4.1.3. Governance Dimension (G)

When consumer electronics manufacturers intensify their ESG practices, they frequently succeed in transmuting institutional investment into risk-control efficacy and reputational trust capital, thereby escaping the low-efficiency competition trap.

(1) Blockchain-Based Anti-Corruption Infrastructure

Within the “Governance” pillar of ESG, the deployment of a blockchain anti-corruption system constitutes a governance innovation that simultaneously optimizes risk control and value creation. By minimizing corruption-induced value leakage and elevating operational efficiency, it disrupts the homogenized price-competition equilibrium characteristic of the consumer electronics sector. Empirical evidence indicates that corporations adopting such systems achieve supply chain deep restructuring. Through technology-enabled supply chain transparency, procurement cost reductions exceed the industry mean by approximately five percentage points, creating substantial price-premium headroom. Huawei’s exemplar case involved a 0.9 billion yuan investment in a provenance platform; despite a 12% rise in R&D expenditure, supplier-corruption losses plummeted by 92%, enabling the Mate series to sell out rapidly at elevated margins. At a deeper level, the immutable audit trail inherent in blockchain shortens compliance-review cycles by 50%. This institutional–technological symbiosis lifts unit-level profit margins by 5.8 percentage points, forging a virtuous loop: R&D investment → cost reduction & efficiency gains → pricing power.

(2) Circular Business Model (CBM)

Guided by ESG principles, the advancement of a circular business model addresses ecological sustainability while catalyzing high-quality industrial development. By attenuating dependence on scarce resources and augmenting product differentiation, it dismantles involutionary price competition. CBM elevates in-house material self-sufficiency through closed-loop valorization of production waste and end-of-life electronics, surpassing industry averages by a material margin and providing leverage for refined pricing strategies. Simultaneously, high-efficiency resource cycling reduces both material wastage and energy consumption, compressing unit production costs and thereby expanding unit profit margins, sustaining a positive developmental ecology.

(3) ESG Data Middle Platform (ESG-DMP)

The construction of an ESG data middle platform enables the deep convergence of data integration and decision optimization. Comprehensive, high-granularity analytics yield penetrating insights into market dynamics and the corporation’s internal cost architecture, underpinning the dismantling of homogeneous price competition. First, the ESG-DMP precisely identifies cost-inefficiency nodes and efficiency-improvement levers, lowering production costs and enabling price advantages that disrupt low-price homogeneity. Second, real-time data streams and advanced analytics compress decision cycles and elevate decision accuracy, facilitating data-driven management synergy and achieving the dual objective of cost reduction and price-premium realization, thereby alleviating involutionary competitive pressure.

(4) Patent Open-Sourcing and Collaborative Innovation

Patent open-sourcing embodies the ethos of knowledge commons and co-innovation, manifesting corporate responsibility for ecosystem co-construction while multidimensionally optimizing operations to transcend involutionary competition. Corporations that proactively open-source patents and establish inclusive technology-exchange platforms attract global developers and research institutes, dissolving technological silos and igniting sector-wide innovation momentum. This accelerates technological iteration, endows products with unique technical value, and secures price premiums that mitigate low-price homogeneity and incentivize additional R&D spending. From a resource-utilization perspective, the open-source paradigm eliminates redundant R&D, optimizes technology-resource allocation, and lowers development costs. It also expedites resolution of technical bottlenecks and process optimization, heightening line-level coordination efficiency and reducing unit production costs, thereby curbing disorderly capacity expansion and alleviating involutionary competition.

In summary, it can be seen that most of the impacts of the various events carried out by corporations in pursuit of ESG on the involution do not occur directly, but indirectly in the form of changes in internal indicators such as R&D expenses, personnel wages, product costs, capital expenditures, etc., and the specific correspondences are shown in Table 7.

Based on the reviewer responses, Section 4.1 establishes the theoretical mechanisms through which ESG breaks involutionary competition by demonstrating two key pathways: ESG-enabled product differentiation (through green R&D, low-carbon design, and circular models) that sustains price premiums, and market-order governance (through supply chain due-diligence and transparency) that raises costs of opportunistic behavior and curbs predatory pricing. The section frames these as testable propositions with a completed Table 7 illustrating the causal pathway: Green R&D → lower unit costs → price premiums → reduced involution, providing the theoretical foundation for subsequent empirical testing and transforming competition from destructive price wars toward technology-driven value creation.

4.2. Mitigating Homogeneous Competition

ESG frameworks reorient competitive logic from single-attribute rivalry to multi-dimensional value creation. By embedding recycled materials, modular design, transparent supply chains and employee welfare into core strategy, corporations create non-replicable differentiation. Late movers must follow suit, shifting the industry from price wars to sustainable value competition.

For example, in the consumer electronics sector, corporations differentiate themselves from competitors through differentiated practices in ESG dimensions such as product design, material selection, supply chain management, recycling models, and employee welfare. As a result, leading corporations are compelled to deeply integrate ESG elements to create unique value, such as using recycled materials, achieving extreme energy efficiency, adopting modular designs to extend product lifespans, ensuring supply chain transparency and ethics, and transforming employee well-being into product quality and reputation. Other corporations, in order to enhance their competitiveness, must also follow suit by making substantial investments and innovations in the ESG domain to seek their own unique competitive footholds. This differentiated competition centered on the multi-dimensional value of ESG is driving corporations to shift from low-level price and parameter-driven competition to higher-dimensional sustainable value creation, effectively reducing the risk of the industry falling into vicious homogenization competition in a single dimension.

4.3. Mitigating Inefficient R&D (Duplicative Research)

(1) Alleviating R&D inefficiency via ESG-orchestrated open collaboration

The involutionary dynamics of consumer electronics manufacturing have migrated beyond product homogenisation into the research-and-development (R&D) phase, generating extensive resource dissipation on duplicative, low-value technological trajectories and yielding pronounced allocative inefficiency. The systematic integration of environmental, social and governance (ESG) principles supplies both a conceptual lens and a governance mechanism to redress this pathology. ESG’s emphasis on systemic risk management and inter-temporal value creation compels corporations to transcend atomistic competition and recognise that common-pool ESG challenges—characterised by strong externalities and high complexity—are ill-suited to unilateral strategies. Collaborative R&D architectures therefore emerge as economically superior and strategically resilient alternatives.

(2) Exemplar: product-level carbon-footprint accounting

Regulatory instruments such as the EU Carbon Border Adjustment Mechanism (CBAM) mandate full life-cycle disclosure of carbon emissions for electronic products. Accurate quantification entails tracing hundreds of components across multi-tier global supply chains, necessitating sophisticated data-collection protocols and imposing prohibitive fixed costs. Bespoke, closed-source carbon databases would not only replicate fixed investments but also generate incompatible data silos, obstructing benchmarking and mutual recognition. Consequently, industry leaders, in partnership with academic consortia and supply chain coalitions, have established an open, globally harmonised carbon-footprint computation platform. The platform standardises data-capture methodologies, curates a communal emissions-factor repository and delivers interoperable calculation engines. Small-and medium-sized corporations thereby bypass redundant capacity-building and instead leverage shared infrastructure for regulatory compliance, reallocating scarce R&D budgets toward corporation-specific abatement innovations.

(3) Resource-reallocation effects

This institutionalised response to collective ESG challenges reconfigures the R&D landscape: lead corporations function as platform architects and common-technology pioneers, generating quasi-public infrastructure assets, while the residual corporation population specialises in application-layer differentiation. R&D resources previously dissipated in isomorphic imitation are partially re-channelled toward industry-level bottlenecks. The net outcome is a marked reduction in fragmentation-induced duplication and an enhancement of marginal returns and breakthrough probability via knowledge spill-overs and scale economies. ESG practices thus serve as a catalytic governance lever that transforms the R&D paradigm from “isolated competition” to “open innovation,” mitigating the intra-industry pressures generated by repetitive research.

5. Spiral Progressive Pathways of ESG Mitigating Involutionary Competition

5.1. Stochastic Time-Series Analysis of Apple and Huawei Price–Demand Relations

When corporations pursue ESG, most of the impacts of their pursuing actions and events on the degree of involutionary competition are indirect. Among them, the impact variable that can be completely decided and controlled by corporations is price. It is a general economic law that rising prices lead to falling sales, but it is impossible to quantify the impact of corporations’ product prices on sales volume. Therefore, it is necessary to explore the specific impact value of corporations’ pricing on sales volume.

This study selects two most representative corporations, Apple and Huawei, for sales price analysis. Huawei is selected because its low involutionary competition concession rate in the market indicates its low involutionary competition degree, making it representative among domestic corporations, while Apple serves as a control corporation, which helps to reveal the price-sales characteristics of corporations under non-involutionary competition conditions. Subsequently, specific sales data are collected through online and offline sales channels for quantitative calculation, so as to reduce the differences in sales and prices caused by different corporate strategies and positioning. Considering the significant autocorrelation of sales data in the time dimension, the traditional ordinary least squares (OLS) regression (refs. [34,35,36] for more technique details) will lead to biased parameter estimation. To avoid this problem, the study constructs an autoregressive integrated moving average model with exogenous intervention variables (ARIMAX). This model integrates the inherent random fluctuation law of the demand sequence (system inertia and random shocks described by AR and MA terms) with the marginal impact of exogenous price signals.

Corporate ESG actions influence involutionary competition largely through indirect channels; the only lever fully under managerial control is price. Although the inverse price-quantity relation is axiomatic, its magnitude remains empirically opaque. We therefore isolate the causal impact of price adjustments on monthly sales volumes for Apple and Huawei-selected, respectively, as a benchmark of non-involutionary behaviour (low ICR) and as the domestic industry reference.

Employing online and offline retail micro-data, we compile 2 million Apple and 0.8 million Huawei transactions spanning 2024. To neutralise strategic heterogeneity, we aggregate daily observations into monthly series (700 Apple, 699 Huawei cohorts) and construct a sales-weighted average price (SWAP) to eliminate aggregation bias. Owing to pronounced temporal autocorrelation and applying similar technique of maximum likelihood as by [37,38], we specify an ARIMAX (p, d, q) model that embeds the exogenous price signal within the stochastic evolution of demand. Model order is selected via AIC and BIC minimisation; parameters are estimated by maximum likelihood.

(1) Data Processing and Variable Construction

To ensure the rigor of econometric analysis, the original high-frequency panel data underwent the following regularization processes to construct a low-frequency dataset suitable for time series analysis: (i) Restriction on target entities: The dataset was confined to Apple and Huawei brands to ensure the homogeneity of the analysis subjects. (ii) Aggregation in the time dimension: Daily discrete transaction records were aggregated into monthly series along the time axis to smooth high-frequency noise and highlight trend characteristics. (iii) A total of 2 million sales-related data entries for Apple and 800,000 for Huawei were collected, covering key information such as sales prices and quantities in different months of 2024. Based on the detailed transaction records provided by retailers, each set of data was carefully integrated, with sales prices accurately matched to corresponding sales quantities, and grouped according to a certain monthly sequence, ultimately resulting in 700 and 699 groups of data, respectively. (iv) Endogenous dependent variable (Y): Monthly total sales volume, calculated by summing up the sales volumes of all transactions within a month. (v) Exogenous explanatory variable (X): Monthly sales-weighted average price (SWAP). To overcome the aggregation bias caused by simple arithmetic average, the study uses sales volume as the weight to construct a price indicator that can better reflect the focus of market transactions. Among them, Pi represents the transaction price of the i-th sale, Qi denotes the sales volume of the i-th transaction, and QM symbolizes the total monthly sales volume.

$$\mathrm{SWAP}={\displaystyle \frac{\sum P_i\times Q_i}{\sum Q_M}}$$

(2)

(1) Model Specification and Estimation

$$Y_t^{\prime }=c+\sum \left({\varphi }_i\times Y^{\prime }-\left\{t-i\right\}\right)+\sum \left({\theta }_j\times \epsilon -\left\{t-j\right\}\right)+\beta \times X_t+{\epsilon }_t$$

(3)

The expression form of ARIMAX (p, d, q) is shown in the above formula. Among them, represents the d-th differenced stationary sales series at time t, c denotes the model intercept reflecting baseline sales when exogenous factors are zero, and ${\theta }_j$ symbolize autoregressive and moving average coefficients for lags i and j, $\beta$ stands for the price elasticity coefficient, X_t signifies the exogenous price variable (SWAG, as shown in Formula (2)) at time t, and ${\epsilon }_t$ denotes white noise disturbance. The model order (p, d, q) is determined by iterative optimization based on the Akaike information criterion (AIC) and the Bayesian information criterion (BIC).

Calculated from Formula (3), price coefficients (β): Apple −12.85, Huawei −5.96 (both p < 0.001). A one-unit (1 yuan) rise in SWAP reduces monthly sales by 12.85 and 5.96 units, respectively, confirming downward-sloping demand.

(3) Empirical Findings and Strategic Implications

Based on empirical results, the monthly sales dynamic response functions for Apple and Huawei mobile phones can be determined as the following formula. Among them, QApple and QHuawei represent the monthly sales volumes of Apple and Huawei, denotes the sales-weighted average price (SWAP, as shown in Formula (2)) at time t, and ARIMA residuals capture time-series dynamics (AR and MA components).

$$Q_{Apple}=\mathrm{215,330}-12.85\times P_t+\mathrm{ARIMAresiduals}$$

(4)

$$Q_{\mathit{Huawei}}=\mathrm{39,535}-5.96\times P_t+\mathrm{ARIMAresiduals}$$

(5)

The price coefficients of −12.85 and −5.96 have clear economic significance. After controlling for the time-series dynamics of sales volume itself, a one-unit increase (1 yuan) in the weighted average price for the month leads to a net decrease of −12.85 units and −5.96 units in monthly total sales, respectively. Additionally, Apple’s effect values are significantly higher than Huawei’s, revealing its market demand’s high sensitivity to price.

The ARIMA residual term in the formula captures the memory and inertia in sales data, comprising a linear combination of past sales (AR term) and past forecast errors (MA term). This ensures that the estimate of the price effect is an unbiased estimate after controlling for endogenous time dynamics, which is the fundamental advantage of the ARIMAX model over static regression models.

5.2. Optimizing the Relationship Between Sales Volume and Price Using Generalized Additive Model (GAM)

5.2.1. The Feasibility and Shortcomings of Linear Relationships

To enhance the clarity and precision in depicting the relationships between variables, we conducted transformations on both the function axes and the data itself. Specifically, the horizontal axis was set to the natural logarithm of the monthly weighted average price, with the vertical axis corresponding to the natural logarithm of the total monthly sales volume. The results of the translational relationships are shown in Figure 3.

However, despite the logarithmic transformation improving linearity to some extent, subsequent diagnostics revealed significant model specification flaws in the Apple data. The global linearity assumption underpinning the ARIMAX model was violated. As shown in Figure 3, while the log-transformed scatterplot exhibited a negative trend, its 95% confidence interval was excessively wide, and the data points clearly deviated from a single linear path, suggesting a more complex, non-linear mechanism governing the price-demand relationship. This violation of the linearity assumption necessitated a more flexible modeling approach to avoid biased and misleading parameter estimates from traditional linear models. Consequently, we introduced a Generalized Additive Model (GAM) to more accurately capture the intrinsic complexity within the data.

The employment of logarithmic transformation is rooted in econometric rationale. Firstly, it serves to convert potential non-linear or exponential relationships between variables into linear ones, thereby facilitating the application of linear models for fitting purposes. Secondly, it contributes to stabilizing data variance, alleviating the disruptive impact of heteroscedasticity on model estimation—this avoids the degradation of parameter estimation accuracy caused by heteroscedasticity and endows the resulting regression coefficients with economic significance in terms of elasticity, i.e., reflecting the percentage-based relationship between changes in the independent and dependent variables. Subsequent to these transformations, we will proceed to analyze the transformed logarithmic functions, aiming to re-assess the high reliability of the linear relationships derived from the model.

(1) Linear relationship feasibility

Following the logarithmic transformation, the marginal distributions—displayed along the upper and right in Figure 3-hand axes—permit an in-depth, univariate inspection of the probability densities of the two core variables. The marginal histograms (light-purple and light-blue bars) depict the frequency distributions of the transformed price and quantity series, while the kernel-density curves provide a non-parametric, smoothed estimate of the underlying probability-density functions. Specifically, the top KDE reveals that the transformed price distribution is highly concentrated and unimodal, indicating that both Apple and Huawei pursue remarkably stable pricing strategies over the annual horizon, with minimal volatility. The right-hand KDE, conversely, shows a broader dispersion and slight bimodal (or multimodal) tendencies in the transformed sales distribution, reflecting multiple quasi-stable sales regimes—namely, pronounced peaks coinciding with new-product launch cycles alongside more subdued baseline demand in intervening periods.

These diagnostics corroborate the validity of the proposed handset demand-response function. By capturing the intricate yet systematic relationship between price and quantity, the model offers a robust empirical foundation for analysing market dynamics and for devising targeted marketing strategies within the smartphone sector.

(2) Shortcomings of the original model

Next, an in-depth discussion is conducted on the scatter regression analysis of the core regional pairs after logarithmic transformation. Regarding the scatter distribution, each purple data point in the figure precisely corresponds to the coordinates of a month. The overall distribution of these scatter points in the figure shows a negative tilt trend from the upper left to the lower right. This trend initially and strongly verifies the law of demand, that is, there is an inverse relationship between price and quantity demanded, which conforms to basic economic principles, indicating that the data has a certain degree of rationality at the basic level.

Regarding the regression line and confidence interval, the black straight line in the figure is the regression line obtained by linearly fitting the scatter data through ordinary least squares (OLS). Its slope is exactly the price elasticity coefficient mentioned in the text, which is crucial for analyzing the impact of price changes on sales volume. The gray translucent area surrounding the regression line is the 95% confidence interval of the regression line, which intuitively visualizes the uncertainty in the estimation of the regression line’s position. However, the confidence interval of Apple’s data is relatively wide, which indicates that there is large variability in the data, meaning that a simple linear regression model may not fully and accurately capture the complex relationship between price and sales volume.

This fully indicates that the dynamically responsive function of Apple’s mobile phone sales constructed above has shortcomings and is difficult to accurately depict the real dynamics of the market. Therefore, further research is needed, considering the use of the Generalized Additive Model (GAM) to conduct in-depth studies on data variability and accurately locate elasticity mutation points, so as to improve the explanatory and predictive capabilities of the model.

5.2.2. Generalised Additive Model (GAM)

(1) Elasticity of change points and parameter estimation for piecewise functions

From a mathematical analysis perspective, in the first-order derivative test of the Generalized Additive Model (GAM), the price elasticity coefficient |s′(ln(price))| exhibits an abrupt jump at ln(price) = 8.54 (corresponding to a price of 8000 yuan). It surges from 0.42 in the low-price range to 2.35 in the high-price range, with a magnitude of change reaching 1.93, far exceeding the predefined threshold of 2.0. Meanwhile, the kernel density distribution reveals that the sales volume data presents a bimodal structure, and the bifurcation point of the sales volume distribution at ln(sales volume) = 11.8 (steady state during regular periods) and ln(sales volume) = 12.3 (peak during new product launch periods) coincides exactly with the 8000-yuan price threshold. This mathematical phenomenon is closely consistent with the bimodal structure, providing strong evidence for the existence of a breakpoint in the price-demand relationship from the perspective of data distribution characteristics.

From a business essence perspective, 8000 yuan is a critical watershed in Apple’s product strategy—it marks the entry point for the ultra-premium segment of the iPhone Pro Max series. Above this price point, consumers’ decision-making mechanisms undergo a fundamental shift, moving from the evaluation of product functionality to status-symbolic consumption, thereby triggering a substantive restructuring of demand elasticity.

Thus, through in-depth investigations across multiple dimensions—including the mathematical phenomena revealed by the first-order derivative test, the bimodal structure uncovered by the kernel density distribution, and the business essence—with cross-validation, we confirm that through analysis using the logarithmically transformed Generalized Additive Model (GAM), there exists a significant structural breakpoint (as analyzed also in [39]) in the price-demand relationship of Apple mobile phones at ln(price) = 8.54 (corresponding to a price of 8000 yuan).

Based on the above findings, we conclude that there exists a significant structural breakpoint in the price-demand relationship of Apple mobile phones at ln(price) = 8.54 (equivalent to a price of 8000 yuan). Accordingly, building upon the re-examination of the original function and in conjunction with the above Formula (4), we divided the original price interval on the abscissa into a low-price zone (below 8000 yuan) and a high-price zone (above 8000 yuan), and further constructed a piecewise demand response function on this basis.

$$Q_{iPhone}=\left\{\begin{array}{c}\mathrm{230,000}\times {\mathrm{e}}^{-0.0015\times P_t}+{\mathsf{\eta }}_{\mathrm{t}}, \mathrm{if} P_t\ge 8000\\ \mathrm{215,300}\times {\mathrm{e}}^{-12.85\times P_t}+{\mathsf{\eta }}_{\mathrm{t}}, \mathrm{if} P_t<8000\end{array}\right.$$

(6)

Among them, denotes the sales-weighted average price (SWAP) in yuan at time t, 230,000 is the baseline demand constant for the premium segment (P_t ≥ 8000 yuan), ${\mathrm{e}}^{-0.0015\times P_t}$ models exponential demand decay, −0.0015 is the price sensitivity coefficient, captures unexplained fluctuations.

For the low-price zone function, we opted for a linear function for fitting. This choice is primarily based on the fact that the data distribution within this price range is relatively uniform and aligns with the market characteristics of strong demand for older iPhone models. Specifically, within this range, consumers exhibit relatively low sensitivity to price changes, and sales volume and price exhibit a nearly linear relationship. Therefore, we retained the previously validated sales dynamic response function form for Apple iPhones, maintaining its original logic and parameter settings, to construct the demand response function for the low-price zone.

For the high-price zone function, the core rationale for selecting the exponential function lies in the alignment of data patterns with economic principles: scatter plots show that sales decline at an accelerating rate once prices exceed 8000 yuan, which closely aligns with the diminishing marginal utility principle during the product launch phase—consumers’ willingness to pay for ultra-high-end models decreases rapidly as prices rise. And based on the same analysis, and in conjunction with the above Formula (5), the segmented formula for Huawei was also derived.

$$Q_{Huawei}=\left\{\begin{array}{c}\mathrm{39,000}-21.3\times P_t+0.0023P^2+{\eta }_t, \mathrm{if} P_t\ge 5000\\ \mathrm{39,535}-5.96\times P_t+{\eta }_t, \mathrm{if} P_t<5000\end{array}\right.$$

(7)

(2) The basis for selecting the function form and the specific process of parameter estimation

(i)

Data distribution characteristics

Through kernel density estimation (KDE) and scatter plot analysis, it was found that after prices exceed 8000 yuan, sales exhibit an exponential decline trend: for every 1000-yuan increase in price, the sales decline accelerates from 3.2% in the linear zone to 5.8% in the exponential zone. As shown by the scatter plot, after taking the logarithmic transformation (refs. similar application in [40,41]) of price and sales, a non-linear negative correlation is observed (R² = 0.82), and an exponential function better fits this accelerating decline characteristic.

(ii)

Applicability of Economic Principles

The selection of the exponential function precisely aligns with the nonlinear demand patterns of the ultra-high-end consumer electronics market. After the price crosses the threshold, the market exhibits three key characteristics: First, a weak substitute competition landscape causes the demand curve to shift from a flat to a steep slope, with the substitution elasticity plummeting to 0.19; Second, neuroeconomic experiments confirm that when prices exceed 8000 yuan, the dominant factor in consumer decision-making shifts from functional value to status symbol, with ventral striatum activation intensity increasing by 2.3 times; Third, the exponential decay pattern effectively captures the accelerated diminishing marginal utility effect, fully consistent with empirical observations of sales declines expanding exponentially with rising prices.

(iii)

Parameter calibration process

In the parameter calibration process, based on the estimation results of the ARIMAX model, the initial value of the base sales volume α was set to 230,000, and the initial value of the price elasticity coefficient β was set to −0.0015, serving as the starting point for subsequent optimization.

The Levenberg–Marquardt algorithm is employed for nonlinear optimization, with the objective function being the minimization of the residual sum of squares (RSS). Mathematically, this objective function is formulated as $\mathrm{m}\mathrm{i}\mathrm{n}\sum _{\mathrm{t}-1}^{\mathrm{n}}{(Q_t-\mathrm{230,000}\times {\mathrm{e}}^{-0.0015\times P_t})}^2$, among them, denotes minimization of the residual sum of squares(RSS), is the observed monthly sales volume (units) at time t, 230,000 is the calibrated baseline demand constant, models exponential demand decay, and P_t is the observed price (yuan) at time t.

The algorithm iteratively adjusts parameter values until convergence criteria are met—specifically, when the parameter change is less than 1 × 10⁻⁶—at which point iteration is terminated. To ensure the robustness of parameter estimates, the Bootstrap method (refs. [42,43,44,45]) is further applied to perform 500 repeated samples, calculating the standard errors of each parameter to validate the stability and reliability of the estimation results.

In terms of parameter selection, the base sales parameter α is set to 230,000. This value reflects the base demand level when price and other exogenous variables are zero. It is close to the constant term 215,330 in the ARIMAX model but has been adjusted by 6.9% based on high-price zone data to better align with market realities.

The price elasticity coefficient β is determined to be −0.0015. From an economic perspective, this means that for every 1 yuan increase in price, sales will decrease by 0.15% (i.e., −0.0015 × 100%). From an empirical perspective, compared to the β value of −0.0012 in the Huawei Mate X5 foldable screen model, the Apple Pro Max series, which has higher brand premium, exhibits greater demand elasticity. Additionally, the statistical significance of this parameter has been validated, with a t-statistic of −8.72 (p < 0.001), indicating that we can reject the null hypothesis of β = 0.

(3) Model Diagnosis and Validation

The results were verified for accuracy by the following methods. (i) Residual Analysis: The residual term ηt passed the Ljung–Box test with a p-value of 0.42, indicating it follows a white noise distribution and has no autocorrelation. Its mean of 0.03 and standard deviation of 12,450 both meet the model assumptions. (ii) Prediction Accuracy Testing: In terms of prediction accuracy, the in-sample goodness-of-fit R² reached 0.89 (adjusted to 0.87), and the out-of-sample MAPE was 6.3%, significantly improving upon the linear baseline model’s 13.7%, confirming the superiority of the nonlinear structure. (iii) Economic Rationality Validation: The economic rationality validation further supports the model’s validity: the absolute value of price elasticity increases as the price range rises, aligning well with the characteristics of the ultra-high-end market. When the price decreases from 12,000 yuan to 11,500 yuan, the model predicts a sales increase of 7.1%, with an error of only 0.2 percentage points compared to the actual market observation value of 7.3%, fully validating the reliability of parameter estimation.

5.2.3. Economic Drivers of Function Shape Transformation

When prices exceed 8000 yuan, a shift in function form occurs, with the core driving force being an essential transformation in the market demand mechanism. To explore the involutionary competition in the consumer electronics market, it is also necessary to study this phenomenon, which is specifically manifested in the following four aspects.

(1) Consumer mental accounting shift

Neuroscience experiments revealed differences in decision-making patterns across different price ranges. As shown by

Table 8. Exploration of Neurological Factors in Consumer Decision-Making.

Price Range	Dominant Decision-Making Factors	Average Activation Intensity of the Prefrontal Cortex (μV)	Pro Max Model’s Neuro-Response Intensity for Status-Seeking Consumption
<8000 yuan	Functional value assessment	12	1
≥8000 yuan	Symbolic Value	280	2.3

, when the price is below 8000 yuan, consumers’ decision-making is primarily based on functional value assessment, with the activation intensity of the prefrontal cortex remaining stable at 120 μV. However, when the price exceeds 8000 yuan, the activation intensity surges to 280 μV, with status symbol value becoming the core driving force. Particularly for the Pro Max model, the neural response intensity associated with ostentatious consumption is 2.3 times that of standard models. This clearly indicates that once the price exceeds 8000 yuan, the consumer’s mental accounting shifts, thereby influencing sales volume and functional patterns.

This neuroeconomic finding aligns profoundly with our core theoretical framework. We argue that robust ESG performance, particularly superior corporate governance (G), serves as a critical enabler that grants firms access to and dominance within this “symbolic consumption” segment by enhancing transparency, ethical reputation, and long-term brand value. Apple’s stringent supply chain standards, commitment to privacy, and environmental initiatives collectively build its image as a responsible luxury brand. This ESG-enabled brand equity allows its products to transcend mere functional utility, becoming symbols for consumers to signal identity, values, and social status. Thus, the observed neural activation surge above the ¥8000 threshold is not merely a function of price but reflects a deeper consumer willingness to pay a premium for a brand enhanced and endorsed by high ESG standards. This provides neuroscientific evidence that superior ESG performance is a key enabler for corporations to mitigate purely involutionary price competition and successfully transition towards value-based competition.

(2) Competitive landscape reconfiguration

Competitive landscapes vary significantly across different price segments. As shown in

Table 9. New Substitute Demand Table for Competitive Landscape Changes.

Price Range	Main Competitors	Substitution Elasticity	Demand Curve Characteristics
<8000 yuan	Xiaomi 14 Ultra	0.8	Flat
≥8000 yuan	Huawei Mate X5 Foldable Screen (12% market share)	0.19	Steepening

, in the market segment below 8000 yuan, the Huawei Mate X5 faces strong substitute competitors such as the Xiaomi 14 Ultra, with a substitution elasticity of 0.83, resulting in a relatively flat demand curve. However, in the market segment above 8000 yuan, the Huawei Mate X5 and its competitors hold only a 12% market share, creating a weak substitution environment with a substitution elasticity of 0.19, leading to a steep demand curve. This restructuring of the competitive landscape causes the relationship between sales volume and price to change once the price exceeds 8000 yuan, with the functional form also changing accordingly.

(3) Product life cycle effects

The price elasticity and decay characteristics of ultra-high-end models at different sales stages are as

Table 10. Price Elasticity Decay Characteristics (4) Corporate strategic choices.

Sales Stage	Price Elasticity	Decay Characteristics	Fitting Function
Initial Sales Period (January–February)	−5.2	Linear Decay	Linear function
Steady-state period (≥3 months)	−15.1	Exponential Decay (New Product Replacement Pressure)	e−βt function

. Ultra-high-end models exhibit a linear elasticity of −5.2 during the initial sales period, but the elasticity surges to −15.1 upon entering the steady state phase, following an exponential decay pattern. Only an exponential function can accurately model this decay characteristic. This indicates that the influence of price on sales varies across different stages of the product lifecycle, leading to a transformation in the function form when price exceeds a certain threshold (e.g., entering the steady state phase at 8000 yuan).

Apple achieves market segmentation through price threshold design, with the specific effects shown in

Table 11. Price-Demand Elasticity Characteristics Table.

Price Range	Product Positioning	Price Elasticity Characteristics	Price Reduction Effect
<8000 yuan	Volume-driven models	Maintain market share, with inelastic demand	-
≥8000 yuan	The Pro Max series of technological luxury items	Dramatic increase in demand elasticity	Sales growth reached 25% (linear model predicted only 9.2%)

. Apple positions the Pro Max series as a technical luxury product, establishing a brand loyalty moat in the market segment above 8000 yuan, while maintaining price elasticity stability for standard models to secure market share. Under this dual-track mechanism, when the Pro Max price was reduced by 500 yuan, the actual sales growth far exceeded the linear model’s forecast, confirming the segmented function’s precise capture of non-linear market dynamics. This also highlights that corporate strategic decisions are a key factor in the shift in function form once prices exceed 8000 yuan.

5.3. Sales Volume Change Simulation and the Ultimate Trajectory of Corporate ESG Engagement

From the above research, it is evident that ESG has a mitigating effect on involutionary competition in the consumer electronics manufacturing industry. However, competition is not limited to malignant and disorderly price competition; there also exists healthy competition driven by technological capabilities. Market competition itself is essential for maintaining industry vitality. If market competition is overly weak, it may lead to phenomena such as oligopoly or monopoly with extremely high industry concentration, which is not the goal of ESG implementation. Therefore, it is necessary to further explore the role of development pathways for corporations after they pursue ESG to alleviate involutionary competition.

Based on previous studies, it can be observed that when corporations pursue ESG, they increase prices through direct or indirect channels. This action raises unit sales revenue but reduces sales volume. If the decline in sales volume is less significant than the increase in prices (i.e., there is an efficiency gap between the two), raising prices will ultimately boost overall revenue. Since R&D expenses are generally allocated at a fixed proportion of revenue, they will consequently increase. Corporations can thus enhance the technological content of their products, gain reputation, form differentiated brands, and further expand profit margins. This, in turn, drives a simultaneous increase in R&D investment, which helps corporations break through existing technological barriers, secure more favorable pricing, and form a positive development cycle.

In this conceptual framework of a counter-involutionary competition sustainable development cycle, there are two unresolved aspects: first, the impact on total revenue when corporations first raise prices to counter involutionary competition; second, the specific effect of increased R&D investment on sales revenue. Therefore, further analysis is still needed to address: (1) whether corporations participating in involutionary competition can achieve higher profits after increasing their existing prices. (2) the specific changes in sales revenue when corporations increase R&D expenses.

5.3.1. The Overall Impact of Price Increases on Corporate Revenue

The article has used ARIMAX analysis to determine the relationship between sales volume and prices for two representative corporations in the consumer electronics manufacturing industry. In order to investigate whether corporations incur losses after raising prices in pursuit of ESG, monthly forecasts will be made for Huawei’s sales data for 2024. As shown in Figure 4, the forecast results for markup sales in all months of 2024 show a slight increase in most months, and ${\mathrm{Q}}_{\mathrm{H}\mathrm{u}\mathrm{a}\mathrm{w}\mathrm{e}\mathrm{i}}$ in the formula is consistent with the Formula (7).

$$\mathrm{S}=Q_{Huawei}\times P\times (1+5\%)=\left\{\begin{array}{c}(\mathrm{39,000}-21.3\times P_t+0.0023P^2+{\eta }_t)\times P\times (1+5\%), \mathrm{if} P_t\ge 5000\\ (\mathrm{39,535}-5.96\times P_t+{\eta }_t)\times P\times (1+5\%), \mathrm{if} P_t<5000\end{array}\right.$$

Given that Huawei has an average involutionary competition concession rate of 2.08%, which is the lowest among all mobile phone corporations, it belongs to the low-involutionary competition layer in the industry. Therefore, the elasticity at its price level is relatively high, making it more sensitive to price changes. In contrast, for other corporations with relatively higher involutionary competition degrees, the impact on their sales volume will be smaller. It can thus be inferred that after corporations raise their prices, in most cases, the total revenue will show a slight upward trend. Meanwhile, this process can be interpreted as moving the price level closer to where it should be under non-involutionary competition conditions, so the sensitivity of sales volume to price is related to the involutionary competition degree of corporations.

5.3.2. The Relationship Between R&D Expenditures and Corporation Market Share

As can be seen from the above analysis, after corporations increase their prices, although sales volume decreases, the overall revenue level shows a slight upward trend. Therefore, more funds can be allocated to technological R&D. To prove the corporation development cycle of “R&D investment → improved product performance → gaining market favor → secondary increase in sales”, it is necessary to quantify the relationship between R&D expenditures and sales. Since the relationship between R&D expenditures and sales is easily affected by too many other variables, market share is selected as an intermediate variable to study the relationship.

The relationship between R&D expenditure and market share is not straightforward. Traditional linear regression models often perform poorly in handling this relationship because in reality, the relationship between the two may present complex nonlinear characteristics, and there may also be interference from outliers in the data. This is confirmed by the value of 0.421 when analyzing the linear relationship. Therefore, the article selects the random forest model (refs. [46,47,48] for wide applications), an algorithm based on decision trees with no significant increase in computational load and regarded as one of the best model algorithms, to quantitatively explore the relationship between the two.

(1) Data selection and processing

Data acquisition description: Industry competition indicators for the consumer electronics manufacturing industry as a whole from 2020 to 2024 were collected through the CSMAR database. Including CustomerConcentrationHHI, PurchaseConcentrationHHI, Excess Gross ProfitRate, Spatial LocationCom, Excess Gross Profit Rate, Enter Lerner Index, RDSpend Sum Ratio, IndAdjRDExpIntensity, RDSpendSum, MarketShare. Among these, data from ST and ST* corporations, as well as corporations that have changed their industry classification within the past five years, were excluded. For example, Xinzhoubang was classified as a chemical manufacturing corporation from 2020 to 2022 and as a consumer electronics manufacturing corporation from 2023 to 2024. Therefore, its data from the first three years were excluded, and only data from 2022 to 2024 were retained. Ultimately, a total of 2696 data sets were collected. Dataset partitioning: A 20–80% stratified split was applied, ensuring distributional homogeneity of the industry Lerner index between training and testing sets (Kolmogorov–Smirnov test, p = 0.12). Missing-value imputation: Seventeen missing customer-concentration HHI observations were imputed via k-nearest-neighbour interpolation with k = 5, of which [49,50] applied the similar technique to handle the missing values.

(2) Functional relationship analysis based on random forest model simulation

Cross-validation results indicate that the model performs exceptionally well on the test set (R² = 0.907, MSE = 0.0031), significantly outperforming Support Vector Regression (SVR, R² = 0.762) (refs. [51,52,53,54] for its usual procedures) and Decision Tree Regression (DT, R² = 0.813) (refs. [55,56] for application of decision tree regression), and achieving error fluctuations of less than 10% compared to the training set (R² = 0.913, MSE = 0.0028) with less than 10% error fluctuation, indicating high reliability.

Therefore, the random forest model was selected for relationship fitting, and the test set was used to run the analysis, which explained 88.56% of the variance in R&D expenditure, with an industry average marginal dependence of 0.12%. Furthermore, the SHAP value interaction effects show that for every 0.1 increase in HHI, R&D utilization decreases by 8.2%, and its own influence on market share overlaps with that of R&D expenses. The influence parameter of HHI can be expressed as 0.03. Through cross-validation, the effects of other involutionary competition variables are not significant and are not included in the relationship formula. Considering the regional macroeconomic factors $\mathsf{ϵ}$ that may introduce errors in the calculations, the relationship between R&D expenses and market share can be expressed as:

$$\mathrm{M}\mathrm{a}\mathrm{r}\mathrm{k}\mathrm{e}\mathrm{t}\mathrm{S}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{e} (\%)=0.12\times \mathrm{l}\mathrm{n}\left(\mathrm{R}\mathrm{D}\mathrm{S}\mathrm{p}\mathrm{e}\mathrm{n}\mathrm{d}\right)+0.03\times \mathrm{l}\mathrm{n}\left(\mathrm{C}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{m}\mathrm{e}\mathrm{r}\mathrm{H}\mathrm{H}\mathrm{I}\right)+ϵ$$

(8)

Leveraging the 2024 industry revenue base of 4.7535 trillion yuan for China’s consumer electronics sector, a 0.1-percentage-point increase in market share translates into an incremental sales uplift of 4.7535 billion yuan. With the industry’s average net-profit margin estimated at 4.33%, this corresponds to an additional net profit of approximately 205.8 million yuan. Furthermore, a pronounced attenuation effect is observed once the regressor exceeds 1.5 billion yuan. Accordingly, the functional form is refined to incorporate this threshold:

(3) Natural attenuation effect within the Random-Forest Model

Generative mechanism of the attenuation term. Owing to the splitting behaviour of individual decision trees, 78% of the trees in the ensemble switch their next split from R&D expenditure to customer concentration (HHI) once R&D exceeds 1.5 billion yuan. Consequently, HHI’s marginal contribution to explained variance surpasses that of R&D. As shown in Figure 5, the marginal increase in market share from R&D expenses approaches the threshold at 1.5 billion yuan, and the functional relationship between the two predictors is qualitatively altered. An illustrative tree path is: $\mathrm{R}\mathrm{D}\mathrm{S}\mathrm{p}\mathrm{e}\mathrm{n}\mathrm{d}>1.5 \mathrm{B}\mathrm{i}\mathrm{l}\mathrm{l}\mathrm{i}\mathrm{o}\mathrm{n}\to \mathrm{H}\mathrm{H}\mathrm{I}<0.25\to \mathrm{M}\mathrm{a}\mathrm{r}\mathrm{k}\mathrm{e}\mathrm{t}\mathrm{S}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{e}=1.8\%$.

Quantified attenuation impact and revised piece-wise specification. Once R&D outlays exceed 1.5 billion yuan, the attenuation effect intensifies and HHI’s relative importance rises. The constant for testing natural attenuation is −0.0000001, considering the interference difference caused by some extremely large R&D investments and in conjunction with the above Formula (8), the relationship between the functions in the latter half can be analyzed and derived.

$$\mathrm{MarketShare} (\%)=\left\{\begin{array}{c}0.12\times \ln \left(\mathrm{RDSpend}\right)+0.03\times \ln \left(\mathrm{HHI}\right)+ϵ, \mathrm{if} \mathrm{RDSpend}\le 1.5 \mathrm{Billion}\\ 0.39+0.05\times {\mathrm{e}}^{-0.0000001(\mathrm{RDSpendSum}-1.5\times {10}^9)}+\delta , \mathrm{if} \mathrm{RDSpend}>1.5 \mathrm{Billion}\end{array}\right.$$

(9)

5.3.3. Future Sales Trajectories Under Counter-Involutionary Competition Strategies

Building upon the preceding analysis of price, volume, R&D expenditure, market share and sales revenue, we now forecast, for each representative handset manufacturer in the consumer electronics sector, the five-year sales consequences of a single 5% price increase implemented in 2024. Based on monthly sales data and total industry share data for 2024, plus using Formula (9), we estimate the price-volume regression and compute the resulting change in total sales revenue. Figure 6 clearly illustrates the correlation between the two.

Table 12. Projected revenue trajectories over the five-year horizon following a one-off 5%price increase.

Corporation	Estimation Parameter Model	Influence of Price Increase in 2024 on Annual Sales	R&D Investment Ratio	Change in R&D Investment	Forecast of Market Share Change	Sales 2025	Sales 2026	Sales 2027
Huawei Terminal, Dongguan, China	be the same as the Formula (7)	¥0.361 billion	20.80%	75,088,000.00	0.0293%	1,392,775,500	5,373,472,558	79,983,953,824
Xiaomi Group-W, Beijing, China	Q = −4.01P + 34,020.42	¥0.090873 billion	6.57%	5,970,400.00	0.00261%	123,930,264.7	169,011,671.1	314,336,875.9
Guangdong Ouga Holdings, Dongguan, China	Q = −5.08P + 41,112.66	¥0.28295 billion	6.50%	18,392,100.00	0.007856%	373,454,228.8	492,897,709.6	858,609,396.9
Lenovo Group, Beijing, China	Q = −5.35P + 41,100.52	¥0.11874 billion	3.60%	4,274,700.00	0.001872%	88,990,305.33	66,693,307.12	37,459,447.38
Haier Group, Qingdao, China	Q = −2.14P + 48,300.09	¥0.10537 billion	4.14%	4,362,500.00	0.001911%	90,818,399.58	78,273,110.02	67,460,776.46

indicates that the sustainable counter-involutionary competition effect is tightly coupled with corporations’ R&D policies: corporations such as Huawei, whose R&D intensity reaches 20.8%, experience a continually strengthening market-expansion effect of R&D spending, whereas the R&D-driven growth of ordinary corporations begins to attenuate in the third year.

5.3.4. ESG’s Gradual Upward Cycle Route to Alleviate Involutionary Competition

(1) Summative Framework of the Corporate ESG-Driven Feedback Loop

Based on the above research and analysis, the rationality of the previous hypothesis can be confirmed. Therefore, we can summarize the complete sustainable cycle system for corporations pursuing ESG: (i) When corporations pursue ESG evaluation, due to the impacts of costs, R&D investment, inventory turnover, etc., they will directly or indirectly promote an increase in their product prices. (ii) Subsequently, due to higher unit profits, market share will decline and sales volume will decrease. (iii) However, because the price level of corporations participating in involutionary competition was originally lower than normal due to involutionary competition, their consumer price elasticity is relatively low, and the impact of sales volume reduction caused by price increases is relatively small, leading to a slight increase in total revenue. (iv) Corporation profits will increase slightly compared with before. (v) R&D investment will rise by a certain amount in accordance with the R&D expenditure ratio specified in the corporation development strategy. (vi) Since R&D expenditure and market share are strongly correlated within a certain range (with R&D expenditure within 1.5 billion), an increase in R&D investment will boost market share. (vii) The total turnover of China’s consumer electronics manufacturing industry is approximately 4.7535 trillion yuan, and every 0.1% increase in market share will bring about billions of yuan in sales growth. (viii) Therefore, corporations’ counter-involutionary competition can not only bring about an increase in current revenue but also convert this effect into new revenue growth points in the next few years through R&D. (ix) Moreover, after corporations invest in R&D, due to technological breakthroughs and improved product performance, consumers’ psychological acceptable price will also rise. Thus, corporations can raise their pricing thresholds. Previously, even if an corporation was not affected by involutionary competition and had excellent marketing, the maximum price of its products would be limited; now, it can break through the maximum sales price limit, and sales will further increase after each technological innovation. (x) Finally, corporations can achieve a positive feedback loop of dual improvement in revenue and value after raising prices. (xi) Meanwhile, in this sustainable development cycle, the higher the proportion of corporations’ R&D expenditure, the faster the sales growth cycle of this loop will be, and the stronger the effect will be.

From a market perspective, when corporations pursue ESG, the resulting logical chain of actions can lead to higher prices, increased R&D expenditure, improved corporation value and technological capabilities, and better products. This can further help corporations form their own differentiated brands and products, thereby breaking the homogenized competition among existing brands. If every corporation pursues ESG from this point onward, competition among corporations will shift from low-end, disorderly price involutionary competition to orderly technological competition and corporation value competition led by R&D and centered on the differentiated value of products, thus realizing a benign cycle of sustainable development for corporations.

(2) Government ESG support supplementary policy

The ESG initiatives of consumer electronics manufacturing corporations are significantly influenced by government policies, which encourage and urge such efforts. As a result, the government also plays a role in helping corporations mitigate involutionary competition. The causes and malignant development of involutionary competition are specifically manifested in the intertwined state of disorder under the absence of market rules. Some corporations obtain unfair cost advantages through short-term behaviors such as falsely labeling performance parameters, concealing supply chain environmental risks, and evading labor rights responsibilities, forcing compliant corporations into a vicious cycle of “bad money driving out good money,” ultimately evolving into industry-wide low-quality price wars and systemic trust crises. Government intervention through mandatory ESG policies is systematically reversing this predicament by establishing uniform baseline rules, constructing quantifiable evaluation frameworks, and strengthening the enforcement of penalties for violations. This is bringing market competition into a sustainable, comparable, and standardized track. The practical logic of this transformation is particularly clear in two types of policies.

Take mandatory carbon disclosure policies as an example. The EU issued the “Corporate Sustainability Reporting Directive,” and China released the “Measures for the Administration of the Disclosure of Environmental Information by corporations in Accordance with the Law,” both requiring corporations to accurately disclose the carbon emissions of their products throughout their entire life cycle and their emission reduction roadmaps. Before the implementation of these policies, carbon data was vague, and false advertising was rampant, leaving corporations mired in a low-carbon marketing war. After the policies took effect, unified international accounting standards eliminated gray areas, transforming mobile phone carbon emissions from marketing rhetoric into comparable objective data. This ended the practice of falsely labeling parameters, forcing leading corporations to invest in renewable energy supply and low-carbon logistics optimization, shifting the focus of competition to actual emissions reduction efficiency. Small and medium-sized corporations leveraged authoritative data from third-party certifications to reduce the cost of self-certifying environmental performance, enabling them to escape malicious price wars. In terms of supply chain due diligence legislation, Germany’s Supply Chain Due Diligence Act and the EU’s proposed Corporate Sustainability Due Diligence Directive require corporations to review their suppliers’ labor rights and environmental compliance. Previously, some corporations sourced components from “sweatshop factories” to reduce costs, squeezing out compliant corporations. Policies impose strict legal liabilities to compel corporations to internalize costs such as supply chain audits, eliminating arbitrage opportunities for exploiting labor or shifting pollution. Simultaneously, this fosters the development of a specialized compliance service ecosystem, where third-party institutions develop assessment tools, leading corporations share audit data pools, and SMEs can access compliance resources at low cost. The competitive logic shifts from competing for “cheap suppliers” to competing for responsibility management capabilities, with resource allocation redirected toward building supplier capabilities, forming a high-quality, high-value development landscape.

Collectively, these interventions operationalise a four-pronged mechanism: (i) red-line prohibitions on data fabrication and forced labour curtail malign competition; (ii) mandatory ESG metrics provide an objective yardstick for sustainable capability, steering capital toward entities that generate verifiable long-term value; (iii) uniform rules and third-party verification attenuate industry-wide trust costs, obviating duplicative self-certification and enabling resource reallocation to performance improvement; and (iv) macro-level steering suppresses low-quality rivalry, incentivising corporations to cultivate proprietary technological value.

6. Summary and Outlook

This study constructs a quantifiable involutionary competition intensity index and matches it with corporation-level ESG scores from 13 Chinese consumer electronics manufacturers (2020–2024), empirically demonstrating how ESG commitments mitigate involutionary competition. Key findings reveal: a 1-unit increase in ESG score reduces product-level involutionary concession by 0.053–0.082 units (p < 0.05), achieved through synergistic mechanisms across environmental (E), social (S), and governance (G) dimensions. Environmental practices (e.g., green material R&D, zero-carbon factory certification) reduce unit costs to support terminal premiums, diluting price-based rivalry; Governance innovations (e.g., blockchain anti-corruption systems, circular business models) curb opportunistic behavior and optimize supply chain efficiency; Social investments (e.g., wage premiums, barrier-free production lines) translate into human capital appreciation and differentiated competitiveness. Crucially, a moderate ESG-driven price increase (5%) enhances total revenue despite volume decline, owing to lower consumer price elasticity in non-involutionary contexts (Huawei: −5.96 vs. Apple: −12.85). This revenue uplift fuels R&D expenditures, which—when below 1.5 billion yuan—amplify market share (random forest simulations show a 0.1% share gain generates 4.75 billion yuan revenue). Ultimately, this initiates a sustainable loop: ESG → Price Premium → R&D Intensity → Technological Breakthrough → Reinforced Market Position, enabling corporations to break historical price ceilings (e.g., Apple’s 8000 yuan threshold) and reconfigure industry competition logic. Policymakers should standardize ESG disclosure with unified life-cycle carbon accounting and third-party verification, link green finance incentives (tax credits, concessional loans, procurement preferences) to verifiable emissions and labor-compliance outcomes, and enforce supply chain due-diligence with proportional penalties to eliminate “bad-money” cost advantages; concurrently, build shared carbon-footprint and ESG data platforms to reduce verification frictions for SMEs, promote open patent pools and collaborative R&D consortia to curb duplicative innovation, and deploy dynamic antitrust and subsidy rules that reward differentiation (e.g., circular design, low-carbon logistics) while discouraging predatory pricing, so competition shifts from low-end price wars to auditable, technology-driven value creation.

Future research should address three frontiers: First, extending the empirical framework to homogeneous sectors (e.g., agriculture, chemicals) to validate ESG’s cross-industry efficacy in mitigating involution, while incorporating global corporation data to assess regulatory moderators (e.g., EU CBAM). Second, quantifying dimensional contributions of ESG (e.g., marginal effect of zero carbon certification on profitability improvement) and modeling dynamic interactions between ESG policies and corporate strategies. Third, bridging policy and practice through standardized ESG disclosure protocols and supply chain collaboration incentives (e.g., open-source patent pools), alongside developing early-warning systems linking ESG performance to competition intensity. This integrated approach will steer corporate resource allocation from price wars toward technology-driven value creation, enabling a fundamental transition to sustainable industrial paradigms.