Does the Low-Carbon City Pilot Promote Household Energy Transition in China?

Abstract

How to promote the household energy transition (HET) has become an important response to extreme climate change. Our paper examines whether a low-carbon city pilot (LCCP) can promote HET. We empirically use the Staggered Difference-in-Differences (DID) model to explore its mechanisms. The results indicate that the LCCP can substantially promote HET. The primary driving mechanism underlying this transition is enhanced governmental emphasis on carbon emission reduction and elevated public environmental awareness. However, the increased local expenditure on energy conservation and environmental protection does not serve as an effective mechanism. The heterogeneity analysis reveals that the LCCP has the most pronounced impact on HET among high-income groups, whereas the effect on low-income groups is relatively minor. Furthermore, the LCCP significantly promotes HET in the eastern region and urban areas, while the central region tends to inhibit it, and the western region and rural areas show no significant effect. The heterogeneity analysis further reveals that the LCCP is effective in Municipalities and Strong-Capital Provinces, where centralized governance and strong political incentives enhance policy implementation. In contrast, the policy shows limited or even negative effects in Non-Municipal Provinces and Non-Strong-Capital Provinces. We provide valuable policy insights for governments to bolster the LCCP implementation to promote HET and achieve carbon neutrality at an earlier stage.

1. Introduction

Nowadays, more and more countries attach great importance to energy conservation and emission reduction. Since the reform and opening up, China’s economy is growing rapidly. However, the extensive development mode results in environmental degradation. Burning large quantities of fossil fuels has led to the excessive release of greenhouse gases like carbon dioxide, further intensifying climate warming. Reducing the reliance on traditional high-carbon energy sources and promoting energy transition is crucial for lowering overall energy consumption and reducing carbon dioxide emissions [1]. While urban and industrial sectors have traditionally been the primary focus in promoting low-carbon development, the household sector is also a significant energy-consuming sector [2]. The International Energy Agency (IEA) declares that the residential sector accounted for 20% of global final energy consumption in 2019 (Data source: https://www.iea.org/data-and-statistics/data-product/energy-efficiency-indicators (accessed on 26 February 2025)). Household energy consumption has become one of the most important sources of global carbon emissions. From 2007 to 2012, China’s total household carbon footprint increased by 19% [3]. Promoting household energy transition (HET) has thus become a critical issue that must be urgently addressed to meet carbon neutrality goals [4,5].

In 2006, China’s total carbon emissions exceeded those of the United States, becoming the largest carbon emitter around the world. In recent years, China has placed significant emphasis on energy conservation and environmental protection, with its economic growth model transitioning from rough and high-speed growth to intensive and high-quality growth. During this period, the Chinese government set carbon emission control targets and implemented a range of environmental protection policies. In 2009, China outlined its 2020 action target for controlling greenhouse gas emissions. The low-carbon city pilot (LCCP) was launched the following year, initially led by five provinces and eight cities. The National Development and Reform Commission (NDRC) later announced the second and third batches of the LCCP list in 2012 and 2017, respectively. Studies indicate that the LCCP has contributed significantly to reducing carbon emissions and conserving energy consumption at the macro-scale [6,7]. However, can this policy promote HET? What specific mechanisms underlie its influence on HET? Few studies have systematically analyzed these questions.

Our paper explores the influence of the LCCP on HET and the underlying mechanisms. The primary data source for our paper is the China Family Panel Studies (CFPS), which is a dataset that encompasses information at individual, household, and community levels. Using the Staggered Difference-in-Differences (DID) model, we estimate the LCCP’s influence on HET in China. The result indicates that the LCCP pushes forward HET significantly. There are two primary mechanisms, which are enhanced governmental emphasis on carbon emission reduction and elevated public environmental awareness. Increased local expenditure on energy conservation and environmental protection has not been proven an effective mechanism. The heterogeneity analyses show that, in terms of different income groups, the policy has a greater impact on high-income groups. From a regional perspective, the influence of the LCCP on HET differs in the eastern, central, and western regions. HET is significantly promoted in the eastern region, while it is inhibited in the central region. In the western region, no significant impact is observed. In terms of urban and rural areas, the impact is significantly pronounced in urban areas and not significantly pronounced in rural areas. The heterogeneity analysis further reveals that the LCCP is effective in Municipalities and Strong-Capital Provinces, where centralized governance and strong political incentives enhance policy implementation. In contrast, the policy shows limited or even negative effects in Non-Municipal Provinces and Non-Strong-Capital Provinces. It reveals that policymakers should account for income, regional, urban–rural, governance, and political incentive disparities when designing similar policies.

Our paper’s marginal contributions are outlined in the following three points: First, previous research has mainly focused on the LCCP’s effects on air quality, energy efficiency, and carbon emissions without comprehensively addressing its influence on HET [7,8]. We investigate the impact in detail and provide an in-depth analysis of its mechanisms. Second, most existing studies on HET emphasize rural regions, examining how factors like income and education influence energy choices but often overlook the role of environmental policies [9,10,11]. We investigate the impact of the LCCP on HET in both urban and rural areas and explore its different effect in these areas. Additionally, it examines the differing effect of the policy on HET in various regions and income groups as well. Third, existing studies about China’s energy transition have mostly operated at the macro-level, mainly at the provincial level [12] and prefectural level [13], examining energy consumption patterns and its influencing factors [6]. However, macro-scale analyses often fail to capture the effects of micro-level variables, which can introduce biases into the findings and diminish the reliability of policy recommendations. To address this limitation, we conduct empirical research at the individual level, incorporating individual and household characteristics into the model. This approach improves model precision and provides a stronger basis for developing more targeted policies. To clarify, we draw on firm-level research conducted by Wen et al., who assess the impact of the LCCP using microdata from A-share listed firms in China’s high energy-consuming industries [14]. Their research applied DID models to analyze the investment and financing mechanisms of these firms, providing important insights into the policy’s impact on corporate performance and financial decision making.

The following sections are organized as follows: Section 2 is policy background, the literature review, and theoretical analysis. Section 3 explains data collection and empirical strategy. Section 4 presents empirical results and analysis. Section 5 shows discussion. Section 6 is conclusions.

2. Background and Theoretical Analysis

2.1. Policy Background

As the largest energy consumer and carbon emitter around the world, China has frequently outlined its energy conservation and emission reduction targets in key meetings and national policy documents. For instance, in 2016, China joined the Paris Climate Agreement, demonstrating its commitment to addressing climate change. In 2020, during the 75th session of the United Nations General Assembly, Chinese leaders announced that China would strengthen its nationally determined contribution, implement more vigorous policies and measures, and take more effective steps to reduce energy consumption. China intends to peak carbon dioxide emissions by 2030 and achieve carbon neutrality by 2060. In 2022, the 20th National Congress of the Communist Party of China stated that promoting green and low-carbon economic and social development is crucial to realizing high-quality development.

To achieve targets for saving energy and reducing emissions, China has implemented a series of targeted policies. In 1995, the State Council issued the 1996–2010 Outline for the Development of New Energy and Renewable Energy, encouraging the development of new energy and renewable energy to optimize the energy structure. In 2007, the State Council issued a notice to establish a unified, comprehensive, and scientific system for energy conservation statistics, monitoring, and assessment, which came into effect in 2008. To implement China’s action targets for controlling greenhouse gas emissions, the NDRC officially launched the LCCP in 2010. Since 2013, China has been running domestic carbon emissions trading pilots in two provinces and five cities. Among these policies, the LCCP has achieved significant results in reducing carbon emissions and conserving energy consumption [6,7]. To date, the LCCP has been implemented in three batches. In 2010, the NDRC issued a notice, initiating the first batch of LCCPs across five provinces and eight cities. In 2012, the NDRC announced the list of the second batch of LCCPs, including Beijing, Shanghai, and Hainan provinces, along with 26 cities (regions) such as Shijiazhuang, Qinhuangdao, and Jincheng. In 2017, the NDRC issued a notice specifying the rollout of the third batch of LCCPs in 45 cities (districts and counties).

The circular on the LCCP explicitly states that local governments are encouraged to explore and implement low-carbon development models that align with local natural conditions, resource endowments, and economic bases. LCCPs primarily achieve low-carbon development through five key measures: formulating low-carbon development and climate change plans, promoting the transformation of traditional industries and the development of low-carbon industries, advocating green and low-carbon lifestyles and consumption patterns, establishing greenhouse gas emission statistics and management systems, and implementing a target responsibility system for controlling greenhouse gas emissions. For instance, Suzhou actively promotes the research and application of clean energy technologies, such as distributed photovoltaic systems, distributed natural gas, and distributed biomass. These initiatives demonstrate that implementing the LCCP has become a crucial factor in promoting HET and achieving the goal of carbon neutrality.

2.2. Literature Review

There are two main strands of the literature closely related to this research. The first focuses on evaluating the effect of low-carbon city policies. Wolff examined the impact of low-emission zone (LEZ) regulation on air pollution in Europe and found that low-emission zones reduced particulate emissions by 9% [15]. Ho et al. integrated low-carbon city concepts into Iskandar Malaysia’s planning, using modeling to simulate sustainable development under a low-carbon city scenario [16]. More relevant to our paper, recent research has examined the influence of the LCCP on energy transition, energy efficiency, and carbon emissions in China [17,18,19]. Lee et al. explored the policy’s impact on energy transition at the prefecture-level city scale. Their findings indicated that the policy significantly accelerated energy transition, with stronger effects in southern, large, non-resource, and old industrial cities [6]. Wang et al. analyzed the LCCP’s effect on energy efficiency and its mechanisms, finding that the policy improved energy efficiency through industrial structure optimization and technological innovation. They also noted variation in the policy’s influence across various types of cities, with a more pronounced effect in large, non-resource, and eastern cities [7]. At the micro-level, fewer studies have explored this topic. Hou et al. examined the policy’s impact on corporate carbon emissions using firm-level tax data. Their results revealed that the policy effectively reduced carbon dioxide emissions by 30% and improved energy efficiency by 36% [8]. Furthermore, in terms of methodology, the existing literature has assessed the effects of LCCPs through qualitative and quantitative methods. Among qualitative approaches, Wang et al. focused on the LCCP of Zhenjiang City and assessed the effect of LCCPs on carbon emissions in Zhenjiang through a case study [20]. As for quantitative methods, Zhang et al. applied the DID model to investigate the impact of LCCPs on carbon emission efficiency in Chinese cities [21]. While case studies provide valuable insights, quantitative methods offer a more precise assessment of LCCP’s impact and establish a clearer causal relationship between the LCCP and HET. Compared with the DID model, the Staggered DID model employed in this paper can assess the dynamic impacts of LCCPs implemented at different times on HET and the cumulative effects of the policy.

The second strand of the literature primarily focuses on energy transition. A segment of this literature explores the factors that influence HET in low-income countries [22,23,24,25]. Nguyen et al. analyzed HET in Vietnam, discovering that region, ethnicity, income, and urban–rural distinctions influence energy transition, with poor and ethnic minority households still largely dependent on traditional energy sources [26]. A study conducted by Emodi et al. revealed that factors such as income, education level, household size, and internet access are significant determinants of HET in Nigeria [27]. Additionally, some of the literature has conducted comparative studies on factors affecting HET across different countries. Pachauri and Jiang conducted a comparative analysis about HET in India and China. They found that the evolutionary trends of household energy use structure and its influencing factors are relatively similar in both countries. In India and China, solid fuel usage exceeds 85% of rural household energy consumption. Due to the low energy efficiency of solid fuels, this leads to higher total energy consumption in rural households compared to urban households. The main factors influencing HET in both countries include urbanization, income, energy prices, energy access, and fuel availability [28]. Heltberg examined household fuel use and the factors driving fuel switching in eight developing countries, including Brazil, Ghana, and Guatemala. The findings showed that the use of traditional biomass fuels is widespread in rural areas, and solid fuels tend not to be displaced even if rural households begin using cleaner fuels. Clean fuel usage exhibited a significant positive correlation with per capita expenditure, household electrification levels, household access to piped water, and education levels [29].

From the literature review, extensive studies have discussed the effects of the LCCP implementation from the macro-scale, as well as studies delving into the factors that influence HET. However, few studies have focused on the LCCP’s influence on HET, and even fewer have conducted an in-depth analysis of its specific mechanisms of action.

2.3. Theoretical Analysis

In response to the dual challenges of climate change and energy shortage, the LCCP targets low-carbon development across various sectors, including industry, transportation, buildings, energy supply, and residential life. When the NDRC released the circular on the LCCP, it clarified the program’s scope and specified the tasks and work requirements to guide local governments in policy implementation. The optimization of energy structures, alongside energy-saving and efficiency improvements, is highlighted in the specific tasks of the three batches of the LCCP, underscoring their importance in policy execution. As a significant energy-consuming sector, the HET is pivotal in optimizing the energy structure. Therefore, local governments would emphasize pushing HET forward when implementing the policy. Furthermore, specific tasks such as planning and constructing energy and heating infrastructure based on low-carbon principles, as well as promoting green and energy-efficient buildings, can offer residents more convenient access to clean energy. It will also contribute to advancing HET. Meanwhile, governments in regions implementing the LCCP encourage enterprises to pursue green technological innovations through tax incentives and financial subsidies. These measures promote the development of low-carbon products, making them more accessible and affordable for residents. Another focus in specific tasks is the active promotion of low-carbon and green lifestyles and consumption patterns. This includes measures such as encouraging demand for low-carbon housing, advocating for low-carbon products, and promoting green and low-carbon modes of transportation like walking, cycling, and public transit. These measures are closely connected to pushing forward HET. Therefore, achieving low-carbon production and consumption will drive the low-carbon transition on both the supply and demand sides of energy [6].

Based on the above analysis, research hypothesis 1 is proposed: the LCCP can push forward HET.

Finance is a crucial tool for the state to promote socio-economic development towards a green and low-carbon transition. Cities that have initiated the LCCP tend to increase their financial investment in energy conservation and environmental protection to support clean energy production, energy conservation, emission reduction, and environmental pollution control. Financial subsidies and tax incentives are often used to encourage enterprises to develop clean energy technologies and related products [30]. These measures not only enable enterprises to achieve energy conservation and emission reduction but also increase the availability of clean energy products across society. Easily accessible clean energy can influence public energy consumption behavior, leading to a shift towards clean energy [31]. Meanwhile, alleviating energy poverty is a significant focus of fiscal expenditure. Energy poverty is a critical barrier to HET, as it refers to the condition in which some residents lack access to affordable clean energy and rely on traditional energy sources to meet their needs [26]. Government subsidies for upgrading residential gas pipelines and natural gas pipelines can assist residents in overcoming the problem of not being able to use clean energy due to outdated equipment or high replacement costs. Direct subsidies for natural gas and electricity bills can also encourage residents to shift to cleaner energy consumption by reducing their costs for clean energy.

Based on the above analysis, research hypothesis 2 is proposed: the LCCP can promote HET by increasing local input in energy conservation and environmental protection.

There are two mechanisms for selecting LCCPs. The first list of LCCPs was established through top-down designation by the central government, while the second and third pilot lists were established through a combination of local declarations and expert reviews. Although various selection mechanisms may result in different policy implementation outcomes, the establishment of LCCPs generally leads local governments to increase their focus on carbon emission reduction. This is driven either by coercive pressure from the central government or by local governments’ incentive to improve their performance metrics [32]. It motivates local governments to adopt various measures to achieve carbon emission reduction. One of the key approaches is requiring enterprises to disclose information about carbon emission and energy consumption [33]. The disclosure of environmental information, such as carbon emission, will reduce information asymmetry among producers, investors, and consumers, thereby encouraging firms to pursue technological advancements. It will also enhance residents’ understanding of energy conservation and emission reduction technologies and related products [34]. In this context, residents may decrease their use of energy-intensive and high-carbon-emitting household energy equipment. This shift will decrease residents’ demand for traditional energy sources and promote a transition to cleaner energy, ultimately leading to HET [4].

Based on the above analysis, research hypothesis 3 is proposed: the LCCP can promote HET by increasing the government’s emphasis on carbon emission reduction.

In contrast to enterprises, it is difficult for governments to impose compulsory energy conservation and emission reduction targets on households. HET largely depends on whether individuals are willing to change their existing energy consumption behaviors. Studies in environmental psychology and sociology indicate that individuals’ energy consumption behavior is heavily influenced by personal environmental values and beliefs. For instance, Hansla et al. found that residents’ positive attitude toward green electricity is associated with a higher willingness to pay for green electricity in Sweden [35]. Weber and Perrels quantified the impact of lifestyle on current and future energy demand by constructing a model to simulate various scenarios. They found that social values and lifestyle choices significantly influence residents’ energy consumption behavior [36]. Cities implementing the LCCP often encourage residents to adopt green and low-carbon lifestyles by organizing educational activities on energy conservation and emission reduction, displaying relevant posters, and distributing pamphlets. Additionally, governments in LCCP-implementing regions can promote the use of low-carbon and environmentally friendly products, such as energy-efficient air conditioners and refrigerators, by offering subsidies and other incentives. They can also reduce the use of disposable plastic bags and excessive packaging through regulatory measures and public awareness campaigns. Furthermore, providing residents with designated waste bins and encouraging waste separation can help increase recycling rates and improve resource efficiency. These initiatives can raise residents’ environmental awareness and inspire them to consciously reduce traditional energy use while transitioning to cleaner energy [37].

Based on the above analysis, research hypothesis 4 is proposed: the LCCP can push forward HET by enhancing residents’ environmental awareness.

3. Data and Empirical Strategy

3.1. Data Collection

To examine the influence of the LCCP on HET, we combine multiple sets of statistical data, which are presented below:

(1) Household energy consumption data: Household energy consumption is the dependent variable, with data from the China Family Panel Studies (CFPS). CFPS is a nationwide, comprehensive social tracking survey project and conducted by Peking University. This dataset is of high quality and has been used in many academic papers [38]. CFPS is biennial, and we utilized data collected in 2010, 2012, 2014, 2016, 2018, and 2020. The questionnaire of CFPS contains questions about energy consumption, and we used the question, “What kind of fuel does your household use most for cooking?” to measure household energy consumption.

Household energy consumption can be categorized into two types: direct energy consumption and indirect energy consumption [39]. Direct energy consumption encompasses household energy use for cooking, heating, cleaning, lighting, transportation, etc. In contrast, indirect energy consumption refers to the energy required for producing goods and services consumed by households [36]. The structure of indirect energy consumption depends largely on the energy sources used in producing these goods and services, mainly influenced by existing production technologies and less by consumers’ choices. Cooking typically represents a significant share of direct household energy consumption [40]. Considering this, we focused on the LCCP’s influence on household direct energy consumption, with a particular emphasis on cooking energy consumption. Figure 1 illustrates the time trend of household energy consumption based on the CFPS dataset. To ensure data accuracy, the dataset was preprocessed as follows. First, the question “Which type of fuel does your household use most for cooking?” was filtered to remove samples with “other” responses or missing values. Second, samples with missing codes for administrative divisions of provinces were excluded.

Several notable trends emerge from Figure 1. The proportion of households relying on firewood and coal for cooking has steadily declined from 2010 to 2020, indicating a gradual transition away from traditional high-carbon energy sources. Simultaneously, the use of bottled gas, LPG, natural gas, and piped gas has increased, becoming the dominant household cooking fuel. This shift reflects improvements in clean energy infrastructure and households’ growing preference for cleaner alternatives. The share of households using electricity for cooking has remained relatively stable, with a slight upward trend, while the adoption of solar energy, biogas, and other alternative sources has remained low. These trends clearly demonstrate household energy consumption structure transitioning from traditional high-carbon sources toward cleaner alternatives in China. This transition aligns with the implementation of the LCCP in 2010, 2013, and 2017. This descriptive evidence supports our point that the LCCP has positively influenced household energy transitions, a relationship further examined through our empirical analysis in the section below.

(2) The LCCP data: According to notices released by NDRC, we collected the data for the three batches of LCCPs from 2010, 2013, and 2017 (The list of the second batch of LCCPs was proposed in December 2012. Considering the delayed effect of policy implementation, this study uses 2013 as the starting year rather than 2012). Given that CFPS data only identify the province where an individual resides, this article integrates the LCCP information at the provincial level. Specifically, we calculated the ratio of the total city area implementing the LCCP in each province to the province’s total area, which serves as the intensity variable for the LCCP implementation in each province. City area data come from the 2020 China administrative division vector data, calculated using ArcGIS 10.8 software. Given the overlapping of the three batches of city pilots, our paper selected the year when the province or the city under its jurisdiction first launched the LCCP as the year of policy implementation. Figure 2 presents a location map depicting the provincial-scale areas where the LCCP was implemented. The color shades on the bottom panel of Figure 2 indicate the proportion of the total area of LCCPs carried out by each province relative to the total area of the province as of the end of 2017. The color bars, ranging from light to dark, indicate the proportion of new LCCPs in each batch from 2010, 2013, and 2017 to the total area of LCCPs carried out across the three batches.

(3) Datasets for control, instrumental, and mechanism variables: Considering that HET is influenced by individual characteristics, household factors, and regional factors beyond the LCCP, we constructed a multi-scale control variable dataset. Individual and household-level variables such as gender, household size, proportion of elderly in household, and fuel expenditure are derived from CFPS data. Provincial-level data, including the resident population, GDP per capita, accepted invention patent applications, enrollment in general higher education institutions, and the proportions of secondary and tertiary industries, were sourced from China Statistical Yearbook. Total natural gas supply, electricity, and heating consumption were sourced from the China Energy Statistical Yearbook. Divorce rate was sourced from China Civil Affairs Statistical Yearbook. The data sources for the instrumental variables are as follows: The number of employees in the production and supply industries for electricity, gas, and tap water in the manufacturing industry and across all industries were derived from the “1982 China Population Census Data”. Wind speed and height data, necessary to calculate the ventilation coefficient, were obtained from ERA-Interim data, which provide wind speed and mixing height at 10 m heights for a global grid of 75° × 75° cells. The proxy variables for mechanism variables include local fiscal expenditure on energy conservation and environmental protection, the proportion of low-carbon environmental protection word frequency, the total number of letters on environmental complaints, and the haze search index. Data on local fiscal expenditure on energy conservation and environmental protection were from China Statistical Yearbook; the proportion of low-carbon environmental protection word frequency was derived from subword processing of the text in government work reports for each province in China, calculating the proportion of low-carbon environmental protection-related word frequency relative to the total word frequency of the report. The haze search index is the Baidu search index for the keyword “WuMai” (which means haze), obtained from the official website of Baidu Index.

The descriptive statistics are shown in

Table 1. Descriptive statistics.

Variables	N	Mean	SD	Min	Max
Key Variables
HET	205,978	0.315	0.464	0	1
LCCP	205,978	0.374	0.432	0	1
Individual-level Controls
Gender	197,707	0.497	0.500	0	1
Household-level Controls
Household Size	205,978	4.344	1.987	1	26
Elderly Ratio	205,978	15.586	25.840	0	100
Fuel Expenditure(ln)	170,886	3.309	2.029	0.000	8.517
Province-level Controls
Population(ln)	205,978	8.554	0.557	5.829	9.443
PGDP(ln)	205,978	10.670	0.480	9.464	12.009
Secondary Industry Ratio	205,978	42.993	6.901	15.967	60.082
Tertiary Industry Ratio	205,978	47.542	8.286	32.463	83.732
Patent(ln)	205,978	9.814	1.189	5.170	12.285
High Education Population(ln)	205,978	4.580	0.534	1.253	5.518
Total Natural Gas Supply(ln)	205,978	0.892	1.285	−6.908	2.953
Divorce Rate	205,978	2.577	0.866	1.080	5.120
Instrumental Variables
Electricity Gas Employment	203,902	4.884	3.649	1.320	13.995
Manufacturing Employment	203,902	212.111	248.470	35.898	1110.411
Mechanism Variables
Environmental Protection Expenditure	205,978	2.887	0.869	1.319	5.672
Low-Carbon Frequency	205,978	0.289	0.110	0.110	0.724
Haze Search Index	172,356	17.684	12.361	0	100

.

3.2. Empirical Strategy

We focused on testing the impact of the LCCP on HET. In studies assessing policy effects, research subjects are typically divided into an experimental group (regions where the policy is implemented) and a control group (regions where the policy is not implemented). The Staggered DID aims to eliminate systematic differences between experimental and control groups. It achieves this by comparing the difference between the changes in the experimental group before and after the policy implementation with the changes in the control group within the same period. This approach effectively assesses the impact of policy implementation. Since the LCCP was implemented in three batches in 2010, 2013, and 2017, we employed the Staggered DID. Provinces included in the LCCP are classified as the treatment group, while other provinces are placed in the control group. The model is set as follows:

$${\mathrm{HET}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{Treat}}_{\mathrm{pt}}{\ast \mathrm{Post}}_{\mathrm{pt}}+{\mathsf{\beta }}_2{\mathrm{Control}}_{\mathrm{it}}+{\mathsf{\eta }}_{\mathrm{p}}+{\mathsf{\gamma }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

In the formula, the subscripts i, p, and t represent individual, province, and year, respectively. HET_it is the dependent variable of our paper. It is assigned a value of 1 when an individual’s cooking fuel is natural gas/piped gas, solar energy/biogas, or electricity, and a value of 0 when the cooking fuel is firewood, coal, or bottled gas/liquefied petroleum gas (LPG) (Piped gas refers to natural gas delivered through pipelines, a relatively cleaner fossil fuel that emits lower levels of carbon dioxide when burned, thus categorizing it alongside solar/biogas, and electricity. Bottled gas refers to liquefied petroleum gas (LPG) packed into cylinders and containers. Unlike natural gas, bottled gas releases a higher amount of carbon dioxide during combustion, putting it in the same category as traditional fuels like firewood and coal). Treat_pt is a grouping variable; if province p is in the experimental group, Treat_pt takes the value of the total area of cities that have implemented the LCCP in the province as a proportion of the province’s area in year t. If province p is in the control group, Treat_pt is 0. Post_pt is a time dummy variable that is set to 0 before the policy implementation and to 1 after that. The coefficient of the interaction term Treat_pt ∗ Post_pt indicates the net effect of policy implementation. Control_it denotes control variables, which influence household energy consumption structure. η_p is the province fixed effect that represents province-related factors that remain constant over time, while γ_t is the time fixed effect that represents time-related factors, which remain constant across regions.

Table 2. Pre-trend analysis.

Variables	Low-Carbon City Pilot Areas	Non-Policy Implementation Areas	Difference
HET	0.242	0.185	0.057
	(0.428)	(0.388)
Gender	0.499	0.498	0.001
	(0.500)	(0.500)
Household Size	4.493	4.413	0.080
	(1.990)	(1.895)
Elderly Ratio	35.121	37.240	−2.119
	(27.816)	(27.550)
Population(ln)	8.648	8.476	0.172
	(0.457)	(0.582)
PGDP(ln)	10.548	10.431	0.117
	(0.328)	(0.442)
Secondary Industry Ratio	47.828	47.396	0.432
	(3.799)	(5.947)
Tertiary Industry Ratio	42.791	42.350	0.440
	(4.219)	(8.335)
Patent(ln)	9.580	9.295	0.285
	(1.068)	(0.968)
High Education Population(ln)	4.568	4.426	0.142
	(0.438)	(0.536)
Observations	14,373	21,093	35,466

Note: Standard deviations in parenthesis.

presents a comparative analysis of HET and control variables between low-carbon city pilot areas and non-pilot areas in 2012. The selection of 2012 as the baseline year was based on the following two primary considerations. Firstly, data availability constraints limited our analysis, as the CFPS dataset commenced in 2010, with no prior data accessible. Secondly, the initial rollout of LCCP in 2010 had limited geographical scope, encompassing only five provinces and eight cities. The second batch of pilots began in 2013. During the 2010–2013 period, the policy’s effects remained minimal. This can be attributed to two factors, the restricted geographical coverage of the first batch and the inherent temporal lag between policy implementation and practical application. As shown in the table, our paper controlled the pre-trends of these key variables when conducting the Staggered DID estimation.

4. Empirical Results

4.1. Baseline Results

Based on the baseline model constructed above, this section empirically examines the impact of the LCCP on HET.

Table 3. Baseline regression results.

Variables	(1)	(2)	(3)	(4)	(5)
LCCP	0.231 ***	0.235 ***	0.234 ***	0.222 ***	0.225 ***
	(0.069)	(0.067)	(0.068)	(0.066)	(0.071)
Gender		−0.004 **	−0.006 ***	−0.006 ***	−0.008 ***
		(0.002)	(0.002)	(0.002)	(0.002)
HouseholdSize			−0.021 ***	−0.021 ***	−0.021 ***
			(0.003)	(0.003)	(0.004)
ElderlyRatio			−0.001 ***	−0.001 ***	−0.001 ***
			(0.000)	(0.000)	(0.000)
Population(ln)				−0.223	−0.490 *
				(0.211)	(0.285)
PGDP(ln)				0.057	0.102
				(0.151)	(0.216)
Secondary Industry Ratio				0.005	0.010 **
				(0.004)	(0.005)
Tertiary Industry Ratio				0.009	0.017 **
				(0.005)	(0.007)
Patent(ln)				−0.026	−0.022
				(0.032)	(0.033)
High Education Population(ln)				−0.069	−0.150
				(0.097)	(0.129)
Fuel Expenditure(ln)					−0.021 ***
					(0.007)
Total Natural Gas Supply(ln)					−0.014
					(0.023)
Divorce Rate					0.012
					(0.034)
Constant	0.228 ***	0.227 ***	0.335 ***	1.586	3.189
	(0.026)	(0.025)	(0.030)	(2.124)	(2.365)
Province FE	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y
Observations	205,978	197,707	197,707	197,707	163,001
R-squared	0.123	0.126	0.134	0.135	0.115

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

presents the baseline regression results. All models in Table 3 control for the province fixed effect and time fixed effect. Control variables for individual characteristics, household characteristics, and provincial characteristics are added in column (2), (3), and (4), respectively. In column (1), the coefficient for the net effect of policy implementation is significant at the 1% level, indicating that the LCCP significantly promotes HET. The signs of the interaction term coefficients and their significance level remain consistent after adding control variables, further validating the model’s reliability. The estimated result in column (4) shows that the effect of the LCCP on HET is approximately 0.222. This demonstrates that the LCCP implementation leads to a 22.2% increase in the probability of HET. It indicates that the LCCP indeed promotes HET from high-carbon to low-carbon energy sources. Research hypothesis 1 was verified.

The regression results for the control variables suggest that male individuals are less conducive to HET. This can be attributed to the gendered division of labor in many households, where tasks such as fuelwood collection and cooking are predominantly undertaken by women. As a result, women are more directly exposed to the inconveniences of traditional fuel collection and the health hazards of polluting gases generated during cooking, which may motivate them to support HET. In contrast, men, who are less involved in these activities, may lack the same level of incentive to transition to cleaner energy sources. This finding highlights the potential of HET to promote gender equality by reducing the disproportionate burden of energy-related tasks on women [29]. Larger household size is associated with lower likelihoods of HET. This is primarily because larger households have higher energy demands. Switching from traditional to cleaner energy sources can lead to a substantial increase in energy expenditures. Thus, larger households are less favorable to HET than smaller ones. For example, Chai et al. found that household size is positively correlated with household electricity demand, with larger households more likely to experience energy poverty [41]. This suggests that larger households face greater financial pressure during the transition to cleaner energy, making them less likely to adopt clean energy technologies. Similarly, households with a higher proportion of elderly members are less likely to undergo HET. This may be because these households face greater financial pressure to support the elderly and are more likely to reduce their expenditure on cooking fuels, opting for lower-cost traditional energy sources. Fuel expenditure reflects the economic burden of energy consumption, which serves as a proxy for affordability. The coefficient for fuel expenditure is significantly negative, suggesting that higher fuel expenditures are associated with lower levels of HET. This finding aligns with our theoretical expectations, as households with higher energy costs may face greater financial constraints, limiting their ability to invest in cleaner energy technologies.

Furthermore, control variables like the proportion of secondary and tertiary industries, the logarithm of resident population, the logarithm of GDP per capita, the logarithm of the accepted invention patent applications, and the logarithm of the enrollment in general higher education institutions do not have a significant impact on HET. Total natural gas supply serves as a proxy for accessibility to clean energy infrastructure. Natural gas, while not entirely renewable, is a relatively cleaner energy source compared to coal and other traditional fuels. The total supply of natural gas at the regional level reflects the availability of cleaner energy infrastructure. The coefficient for this variable is statistically insignificant, suggesting that the availability of natural gas infrastructure alone does not significantly drive HET in our sample. The possible reason is that the availability of natural gas infrastructure may not directly translate into household adoption, as other factors such as cost and income may play a more significant role. Divorce rate is used to measure social constraints. Divorce rates can indirectly reflect the stability of household structures and social networks, which may influence energy consumption patterns and the adoption of new energy technologies. The coefficient for divorce rate is also statistically insignificant, indicating that social constraints do not have a significant direct impact on HET in our context. This could be because the relationship between social constraints and HET is more complex and mediated by other factors such as income levels.

4.2. Robustness Test

To confirm the validity of the baseline estimation results above, multiple robustness tests are conducted. The results are presented in

Table 4. Robustness test results.

Variables	(1)	(2)	(3)	(4)
	Reduced Tail	Select Sample	Standard DID	Comprehensive HET
LCCP	0.222 ***	0.223 ***		0.084 *
	(0.066)	(0.065)		(0.045)
LCCP_Dummy			0.189 **
			(0.077)
Individual-level controls	Y	Y	Y	Y
Household-level controls	Y	Y	Y	Y
Province-level controls	Y	Y	Y	Y
Constant	Y	Y	Y	Y
Province FE	Y	Y	Y	Y
Year FE	Y	Y	Y	Y
Observations	197,707	158,224	197,707	133,009
R-squared	0.135	0.134	0.135	0.180

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

.

(1) To minimize the possible influence of extreme values on model estimation results, samples with net household income in the lowest and highest 5% are removed. Only samples with net household incomes between the 5th and 95th percentiles are retained for model estimation. The result of the trimmed model is presented in column (1) of Table 3. The estimated coefficient’s magnitude and significance for the interaction term are consistent with those in column (4) of Table 2. It indicates that extreme values do not affect the model estimation results.

(2) To maintain consistent and reliable model estimation results across various subsets of the sample, 80% of the dataset is randomly chosen for model estimation. The random sampling process was designed to yield a representative sample. Column (2) of Table 3 presents the estimation result. The main estimated coefficient is positive and significant at the 1% level, aligning with the baseline estimation result. This consistency demonstrates that the estimated effect of the LCCP is not affected by the sample selection process.

(3) The continuous DID model is replaced with a standard DID model. The treatment group variable Treat_pt is defined based on whether there are cities within a province that have implemented the LCCP. Specifically, the Treat_pt variable is assigned a value of 1 if the cumulative area of cities implementing LCCPs exceeds 50 percent of the province’s total area; otherwise, it is assigned a value of 0. The estimated result is shown in column (3) of Table 3. It reveals that the primary estimated coefficient is significantly positive, despite a slight decrease in its significance level. The robustness test results confirm that the baseline regression result is robust and that the LCCP implementation indeed promotes HET.

(4) To measure HET more comprehensively, we construct an expanded index by incorporating heating, lighting, and appliance use. Since the CFPS data only provide information on cooking fuel and lack variables related to other types of energy consumption, we supplement household-level cooking fuel data with provincial-level electricity and heating consumption. The provincial data, measured in standard coal equivalent, are standardized for comparability. To account for regional differences in heating demand, we assign differentiated weights to cooking fuel, electricity, and heating. For northern provinces, where centralized heating is prevalent, cooking fuel is weighted at 0.7, electricity at 0.1, and heating at 0.2. For southern provinces, cooking fuel remains at 0.7, while electricity is weighted at 0.2 and heating at 0.1. Using this expanded HET measure, we re-estimate the impact of the LCCP on HET. The result is shown in Table 4 column (4). The coefficient of the LCCP remains statistically significant at the 10% level, indicating that the LCCP continues to promote HET even when accounting for a broader range of energy consumption activities, including electricity and heating. This robustness test reinforces the validity of our main findings.

4.3. Endogeneity Test

Reasons such as omitted variables and reverse causality can result in endogeneity problems when estimating the LCCP’s impact on HET. To overcome the potential endogenous bias, the two-stage least squares (2SLS) method is applied in this section to re-estimate the LCCP’s effect on HET.

Drawing from previous studies, the LCCP is considered as an endogenous variable reflecting the technological level of the secondary industry in the historical period. It employs the proportion of the size of employees in the production and supply industries for electricity, gas, and tap water, as well as in the manufacturing industry, relative to the total number of employees across all industries in 1982 (China conducted the third national population census in 1982, yielding detailed statistics on employment categorized by province and industry), to measure the technological level of the secondary industry during that period [42,43]. The underlying logic is as follows: first, the National Bureau of Statistics states that the secondary industry contributes most to China’s carbon dioxide emissions, with manufacturing, power, and heating industries as the primary contributors. Production levels in these industries are significantly positively correlated with carbon dioxide emissions. Carbon dioxide emissions are strongly related to carbon emission policies adopted by local governments. Production levels in these industries correlate with the carbon emission policies adopted by local governments. Second, besides impacting current household energy consumption through their influence on carbon reduction policies, the technological level of the secondary industry from a historical period has virtually no impact on recent HET, nor is it influenced by them.

Since the employment proportion indicators of the two industries mentioned above in 1982 do not change over time, the products of each of these two indicators with the ventilation coefficient are selected as the instrumental variables for the LCCP implementation [44]. The ventilation coefficient is used in this multiplication due to the following considerations: first, it satisfies the correlation assumption. In areas with high ventilation coefficients, ventilation is stronger, allowing carbon dioxide to diffuse more easily, indicating a negative correlation between the ventilation coefficient and carbon dioxide concentration in the area. Thus, the ventilation coefficient correlates with environmental regulations like the LCCP. Second, it satisfies the exogenous assumption. The ventilation coefficient depends on natural factors such as meteorology and is unaffected by economic development or human activities. Based on previous research, the ventilation coefficient is calculated by multiplying the average wind speed with the boundary layer height [45].

Table 5. Endogeneity analysis results.

Variables	First Stage		Second Stage
	(1)	(2)	(3)	(4)
	LCCP	LCCP	HET	HET
Electricity Gas Employment	−0.160 ***
	(0.039)
Manufacturing Employment		−0.004 ***
		(0.000)
LCCP			0.429 ***	0.411 ***
			(0.138)	(0.083)
Individual-level Controls	Y	Y	Y	Y
Household-level Controls	Y	Y	Y	Y
Province-level Controls	Y	Y	Y	Y
Constant	Y	Y	Y	Y
Province FE	Y	Y	Y	Y
Year FE	Y	Y	Y	Y
Observations	195,732	195,732	195,732	195,732
R-squared	0.943	0.948	0.133	0.133
Fstat_first stage	16.53	161.18	-	-

Note: Robust standard errors in parentheses. *** p < 0.01.

shows the estimation results for the instrumental variables. Columns (1)–(2) and (3)–(4) present estimation results for the first and second stages of the 2SLS, respectively. The F-statistics for the first stage for both instrumental variables are well above 10, suggesting that weak instrumental variable tests are successfully passed. What is more important, the second-stage regression results align with the baseline regression estimations above. It demonstrates that the LCCP can promote HET.

4.4. Mechanism Analysis

The results from the baseline model, robustness tests, and endogeneity test indicate that the LCCP indeed pushes forward HET. But what mechanisms enable this policy to be effective? This section aims to examine the mechanisms by which the LCCP promotes HET. It focuses on three key aspects: local input in energy conservation and environmental protection, government’s emphasis on carbon emission reduction, and residents’ environmental awareness.

(1) Local input in energy conservation and environmental protection: The LCCP may influence HET by boosting local input in energy conservation and environmental protection. Governments can achieve this in several ways. On one hand, they might directly improve the energy-saving and environmental protection infrastructure, such as by expanding the coverage of natural gas pipelines. On the other hand, they can encourage households to install solar photovoltaic power generation devices through subsidies or cash incentives. Additionally, governments can support the environmental protection industry by offering special grants, financial subsidies, and other financial tools to promote enterprises that provide resource-saving products. Whether governments directly provide energy-saving service facilities or offer subsidies to encourage households to use energy-saving devices, these efforts can promote HET. To validate this mechanism, the proportion of local fiscal expenditure on energy conservation and environmental protection for each Chinese province from 2010 to 2020 is used as proxy variable for local input in energy conservation and environmental protection. Column (1) of

Table 6. Mechanism analysis results.

Variables	(1)	(2)	(3)
LCCP	0.283 ***	0.125 *	0.167 **
	(0.084)	(0.072)	(0.078)
Environmental Protection Expenditure	0.016
	(0.012)
LCCP × Environmental Protection Expenditure	−0.030
	(0.024)
Low-Carbon Frequency		−0.060
		(0.061)
LCC × Low-Carbon Frequency		0.326 **
		(0.123)
Haze Search Index			−0.003
			(0.002)
LCCP × Haze Search Index			0.004 *
			(0.002)
Individual-level Controls	Y	Y	Y
Household-level Controls	Y	Y	Y
Province-level Controls	Y	Y	Y
Constant	Y	Y	Y
Province FE	Y	Y	Y
Year FE	Y	Y	Y
Observations	197,707	197,707	164,336
R-squared	0.135	0.135	0.109

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

shows the empirical regression result. The coefficient of the interaction term between the LCCP and the mechanism variable shows no statistical significance. The result indicates that LCCP implementation does not affect HET through the mechanism of local input in energy conservation and environmental protection. Research hypothesis 2 failed to verify.

One possible reason for this finding is that local fiscal expenditures on energy conservation and environmental protection may not be effectively targeted or efficiently utilized. For instance, funds allocated for an energy-saving infrastructure might be disproportionately spent on administrative costs or large-scale industrial projects rather than household-level initiatives. Additionally, it could reflect insufficient coordination between local governments and households, leading to a mismatch between policy objectives and actual household demands. For example, local government may invest heavily in expanding natural gas pipelines. However, if it fails to fully understand the actual conditions of rural households, such as income levels and energy usage habits, it may result in many households that are unable to afford the cost of natural gas and prefer to continue using traditional fuels. In this case, despite significant government investment in energy conservation and environmental protection, the policy outcomes cannot reach the expectation.

(2) Government’s emphasis on carbon emission reduction: The LCCP could also contribute to HET by strengthening the government’s emphasis on reducing carbon emissions. Governments of pilot cities tend to prioritize carbon emission reduction, leading to a more extensive disclosure of information on energy consumption and carbon emissions. This increased transparency helps consumers understand the benefits of low-energy products, encouraging them to reduce the use of high-energy products, thus influencing household energy consumption. Additionally, local government departments always lead by example in implementing energy conservation and emission reduction initiatives. It can create a demonstration effect on individual energy consumption behavior and encourage individuals to choose clean energy. To verify this mechanism, we use the proportion of low-carbon environmental protection word frequency in the government work reports from 2010 to 2020 across 31 Chinese provinces as a proxy variable for the government’s emphasis on carbon emission reduction [46]. Words specifically related to low-carbon environmental protection include terms such as “HuangJingBaoHu”, “HuanBao”, “DiTan”, “JianPai”, and “LvSe”, which refer to environmental protection, low carbon, and emission reduction. The estimated result in column (2) of Table 6 verifies this mechanism. It indicates that the LCCP can push forward HET by enhancing government’s emphasis on carbon emission reduction. Research hypothesis 3 is verified.

However, it is important to acknowledge the limitations of using word frequency as a proxy for government emphasis on carbon emission reduction. While word frequency in government work reports reflects the rhetorical focus of policymakers, it may not fully capture the actual implementation or effectiveness of carbon reduction policies. Additionally, the content of government work reports may be influenced by political considerations, potentially leading to an overestimation of the government’s commitment to carbon reduction. Thus, measurement limitations may overestimate LCCP’s promotion of HET through the government’s carbon reduction emphasis mechanism. Future research could complement word frequency analysis with more appropriate measures.

(3) Residents’ environmental awareness: The LCCP might influence HET by raising residents’ environmental awareness. In regions with the LCCP, educational activities are organized to increase residents’ understanding of global warming and energy shortage, fostering greater environmental awareness. As a result, residents might consciously reduce their use of high-carbon fuels and shift to cleaner energy sources [37]. Additionally, in LCCPs, street posters about climate issues and advocating green lifestyles could effectively increase residents’ focus on environmental issues and raise their environmental awareness, encouraging them to promote their household energy consumption toward a more low-carbon lifestyle. To verify this mechanism, the logarithm of a haze search index from 2011 to 2020 (publicly available Baidu Index data only date back to 2011, with no publicly accessible data from prior years) was chosen as proxies for residents’ environmental awareness in each province in China. Columns (3) of Table 6 report the relevant regression results. The coefficient of the interaction term is statistically significant. It suggests that the mechanism through which the LCCP pushes forward HET by enhancing residents’ environmental awareness is confirmed. Research hypothesis 4 is verified.

4.5. Heterogeneity Analysis

4.5.1. Income Disparities

Individuals at different income levels vary significantly in their household energy use, with poorer households relying more on high-carbon-emitting energy sources like coal [3]. Damette et al. showed that, in France, the low-income group was more likely to use fuels such as wood when choosing household energy sources, while the high-income group was more inclined to adopt cleaner energy options [47]. The low-income group was also more likely to experience energy poverty during the energy transition. Therefore, it is vital to understand the LCCP’s various influence on HET across different income groups. In this section, samples are categorized into low-income, middle-income, and high-income groups based on individuals’ income. The three columns in

Table 7. Heterogeneity analysis results: Income disparities.

Variables	(1)	(2)	(3)
	Low Income	Medium Income	High Income
LCCP	0.093 **	0.200 ***	0.256 ***
	(0.043)	(0.056)	(0.091)
Individual-level Controls	Y	Y	Y
Household-level Controls	Y	Y	Y
Province-level Controls	Y	Y	Y
Constant	Y	Y	Y
Province FE	Y	Y	Y
Year FE	Y	Y	Y
Observations	53,402	53,098	52,866
R-squared	0.087	0.115	0.143

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05.

present the estimated impact of the LCCP on HET among these three income groups. The estimated results reveal that the LCCP has a greater influence on HET among the high-income group. In the future, it is important to adopt more targeted measures to promote HET for specific groups, especially for low-income groups.

4.5.2. Regional Disparities

China has a vast territory, which is made of a variety of regions. These regions differ greatly in natural characteristics and economic development. The effect of the LCCP on HET may be different in various regions. In this section, the samples are categorized into the eastern region, central region, and western region [6]. Three columns in

Table 8. Heterogeneity analysis results: Regional disparities.

Variables	(1)	(2)	(3)
	Eastern	Central	Western
LCCP	0.374 ***	−0.318 **	0.235
	(0.055)	(0.102)	(0.305)
Individual-level Controls	Y	Y	Y
Household-level Controls	Y	Y	Y
Province-level Controls	Y	Y	Y
Constant	Y	Y	Y
Province FE	Y	Y	Y
Year FE	Y	Y	Y
Observations	85,792	57,965	53,950
R-squared	0.141	0.117	0.157

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05.

present the estimated influence of the LCCP on HET in these three regions. The results indicate that the LCCP implementation significantly pushes forward HET in the eastern region, which are consistent with the existing findings [6,13].

However, in the central region, the LCCP leads to a reverse energy transition. Emodi et al. noted that Nigerian households experienced a similar reverse energy transition. They thought that decreased household income and low-asset levels in Nigeria influence individuals’ energy consumption choices [27]. In the central region, the adverse effect of the LCCP on HET can be attributed to a combination of economic, industrial structure, policy implementation difficulties, and cultural factors. First, the region’s flat terrain and abundant arable land provide easy access to traditional fuels such as straw, making them a cost-effective alternative to clean energy for households with limited income. Second, the central region’s industrial structure, which is often dominated by agriculture and heavy industries, may create additional barriers to energy transition. For instance, the reliance on traditional fuels in agricultural activities and the high energy intensity of heavy industries could reduce the effectiveness of the LCCP. Third, policy implementation challenges, such as insufficient infrastructure for clean energy distribution and limited local government capacity, may further hinder the adoption of clean energy. Finally, cultural factors, including long-standing reliance on traditional fuels and a limited awareness of clean energy benefits, could also play a role in slowing HET progress.

In the western region, the LCCP does not significantly promote HET. This may be due to the region’s unique geographical and economic conditions, such as its mountainous terrain and lower population density, which increase the cost and complexity of clean energy infrastructure deployment. Additionally, the western region’s relatively underdeveloped economy may limit households’ ability to afford clean energy technologies, even with policy support. Furthermore, the complementary measures of the LCCP may not be well suited to the specific needs of the western region, such as the lack of targeted subsidies for off-grid renewable energy solutions or insufficient technical support for rural households.

4.5.3. Urban–Rural Disparities

As discussed earlier, the effectiveness of the LCCP implementation may be associated with economic development of the region. Economic disparity between urban and rural areas is apparent, which leads to significant differences in energy consumption habits and behaviors among households. In this section, samples are categorized into urban and rural groups based on the individual’s type of settlement. The two columns in

Table 9. Heterogeneity analysis results: Urban–rural disparities.

Variables	(1)	(2)
	Urban	Rural
LCCP	0.157 **	0.027
	(0.071)	(0.048)
Individual-level Controls	Y	Y
Household-level Controls	Y	Y
Province-level Controls	Y	Y
Constant	Y	Y
Province FE	Y	Y
Year FE	Y	Y
Observations	91,327	101,737
R-squared	0.216	0.120

Note: Robust standard errors in parentheses. ** p < 0.05.

display estimates of the impact of the LCCP on HET in urban and rural areas, respectively. The estimated results show that the LCCP in urban areas pushes forward HET significantly, whereas the influence of this policy in rural areas is not significant. A study conducted by Wang et al. found that it is difficult to realize energy transition in rural areas. It indicates that, compared to urban areas, households in rural areas have larger economic pressure when they switch to use clean energy [38]. Furthermore, Tao et al. found that, due to a lack of education, the risks associated with using traditional fuels in rural areas are often overlooked. Lots of individuals choose traditional fuels for cooking and heating. They do not realize that the polluting gases released from burning solid fuels have a large and bad influence on their health [9]. Considering various factors, there is more resistance to realizing energy transition in rural areas than in urban areas. To some extent, these studies can support the finding that the LCCP does not significantly affect HET in rural areas.

4.5.4. Governance Structures and Political Incentive Strategy Disparities

The effectiveness of LCCPs on HET is shaped by governance structures and political incentive strategies. To examine this impact, we group provinces into Municipalities and Non-Municipal Provinces based on governance structures and into Strong-Capital Provinces and Non-Strong-Capital Provinces based on political incentives.

Table 10. Heterogeneity analysis results: governance structures and political incentive strategy disparities.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Municipality	Non-Municipality	Strong Capital (gdp)	Non-Strong Capital (gdp)	Strong Capital (pop)	Non-Strong Capital (pop)
LCCP	0.193 **	−0.271 ***	0.184 **	−0.115	0.248 ***	0.074
	(0.072)	(0.086)	(0.076)	(0.176)	(0.042)	(0.111)
Individual-level Controls	Y	Y	Y	Y	Y	Y
Household-level Controls	Y	Y	Y	Y	Y	Y
Province-level Controls	Y	Y	Y	Y	Y	Y
Constant	Y	Y	Y	Y	Y	Y
Province FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
Observations	68,229	129,478	135,141	62,566	160,664	37,043
R-squared	0.154	0.127	0.153	0.097	0.143	0.120

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05.

displays the impact of LCCPs on HET in different governance structures and political incentive strategies. Municipalities, which operate under centralized governance, show a positive and significant impact of LCCPs on HET. This reflects efficient policy enforcement and resource allocation under centralized control. In contrast, Non-Municipal Provinces exhibit a negative and significant effect, likely due to challenges such as inconsistent enforcement and competing local priorities under decentralized governance.

For political incentives, we define Strong-Capital Provinces as those where the provincial capital accounts for over 50% of the province’s GDP or population. These criteria ensure that our measure of political incentives captures the concentration of economic and demographic resources in the provincial capital, which is often associated with stronger political motivation to achieve policy targets. As is shown in column (3) and (5) of Table 10, Strong-Capital Provinces demonstrate a positive and significant impact of LCCPs, highlighting the role of concentrated resources and strong political incentives in driving HET. In contrast, Non-Strong-Capital Provinces show no significant effect, suggesting that weaker political motivation and resource dispersion hinder policy effectiveness. These findings emphasize the importance of centralized governance and robust political incentives in enhancing LCCPs.

5. Discussion

By using the Staggered DID model and data from multiple sources, we analyze the influence of the LCCP on HET and its mechanisms. The main findings are as follows:

First, the LCCP has significantly contributed to HET in China. As the highest total carbon emitter around the world, China is in the early stages of energy transition. It faces great challenges such as insufficient energy supply and inefficient energy structure. The household is an important energy consumption sector, and its energy transition is particularly crucial. To further promote HET, policymakers should expand the scope of LCCPs and strengthen their enforcement. Specifically, carbon emission reduction targets should be tailored to regional characteristics, such as resource endowments, industrial structures, and energy consumption patterns. For example, regions rich in renewable energy resources could be assigned higher targets for renewable energy adoption, while industrial-heavy regions could focus on energy efficiency improvements and industrial decarbonization. Furthermore, a rigorous carbon emission reduction target assessment mechanism should be implemented to ensure compliance among high-emission households within pilot areas. It may include the regular monitoring and reporting of carbon emissions, coupled with penalties for non-compliance and incentives for exceeding targets. For instance, high-performing households could receive tax breaks, subsidies, or public recognition, while underperforming entities could face fines or restrictions on energy usage. To enhance the effectiveness of LCCPs, policymakers should also prioritize capacity-building and public engagement. Training programs could be established to equip local officials with the technical and administrative skills needed to design and implement energy transition strategies. Public awareness campaigns could be launched to educate households about the benefits of energy transition and encourage behavioral changes, such as adopting energy-efficient appliances or reducing energy waste.

In addition, the baseline regression result reveals that gender (male), larger household size, higher elderly ratio, and lower affordability significantly and negatively impact HET. These findings highlight the importance of considering demographic factors in the design and implementation of LCCPs. To address these barriers, policymakers should adopt targeted interventions that consider the unique challenges faced by these demographic groups. For male-headed households, awareness campaigns and incentive programs should be designed to emphasize the health and economic benefits of clean energy, encouraging greater male participation in energy transition initiatives. For larger households, tiered subsidy programs and energy efficiency measures should be introduced to alleviate the financial burden of higher energy demands and make clean energy more accessible. Households with a higher proportion of elderly members may benefit from tailored financial assistance, such as subsidies for clean energy appliances or low-interest loans, coupled with community-based outreach programs to educate and support elderly individuals in adopting new technologies. Additionally, to address affordability constraints, policymakers should expand clean energy subsidies and low-interest financing options, particularly for low-income households. By implementing these targeted measures, policymakers can ensure that the benefits of HET are equitably distributed and that no demographic group is left behind in the transition to a low-carbon future.

Second, the LCCP promotes HET primarily through enhanced governmental emphasis on carbon emission reduction, along with elevated public environmental awareness. However, the increased local expenditure on energy conservation and environmental protection does not serve as an effective mechanism. Therefore, the government should enhance emphasis on carbon emission reduction and encourage households to adopt clean energy sources. Specifically, it should prioritize carbon emission reduction and improve transparency regarding energy consumption and carbon emissions, enabling people to access policy information more conveniently and comprehensively. It is important to promote green and low-carbon consumption patterns and lifestyles through subsidies and cash incentives as well.

Third, the impact of LCCPs on HET exhibits significant heterogeneity across income, regional, urban–rural, governance, and political incentive divides. (1) The LCCP has contributed to HET across all income groups, with the most pronounced impacts observed in the high-income group and relatively lesser effects in the low-income group. To address this disparity, targeted policy interventions are needed. For high-income groups, the implementation scope of the LCCP should be expanded, leveraging their financial capacity to adopt advanced clean energy technologies. For low-income groups, enhanced clean energy consumption subsidies and low-interest financing options should be introduced to alleviate economic barriers and encourage participation in energy transition. Additionally, community-based programs could be established to provide technical support and education on energy-saving practices tailored to low-income groups.

(2) The LCCP in the eastern region and urban areas significantly promoted HET, while the central region exhibited a trend of inhibition, and the western region and rural areas demonstrated no significant impact. To address these regional disparities, region-specific policy adjustments are essential. In the eastern region and urban areas, where the policy was effective, efforts should focus on scaling up successful initiatives and integrating smart energy management systems to further enhance efficiency. In the western region and rural areas, the LCCP should be strengthened by investing in clean energy infrastructure (e.g., solar panels, biogas systems) and improving energy access to combat energy poverty. Policies should also be focused on reducing energy inequality by ensuring the equitable distribution of resources and subsidies. As for the central region, policymakers should conduct detailed assessments of local energy needs and resource availability to design LCCP interventions that align with local unique characteristics. For example, promoting hybrid energy systems that combine renewable energy with the existing infrastructure could be more effective. Additionally, strengthening local governance capacity and enhancing coordination between central and local governments could improve policy implementation and ensure that LCCP measures are better integrated into regional development plans.

(3) The LCCP policy is effective in Municipalities and Strong-Capital Provinces and shows limited or even negative effects in Non-Municipal Provinces and Non-Strong-Capital Provinces. It highlights the challenges of decentralized governance and weaker political motivation in driving HET. Therefore, policymakers should prioritize strengthening centralized governance in decentralized regions to ensure consistent policy enforcement and resource allocation. Additionally, targeted financial incentives and performance-based rewards should be introduced to motivate local governments in Non-Strong-Capital Provinces to prioritize the LCCP initiatives. For instance, central government transfers could be linked to the achievement of energy transition targets, and local officials could be incentivized through career advancement opportunities tied to successful LCCP implementation.

Additionally, it is important to consider the potential trade-offs of increasing clean energy consumption subsidies. While subsidies can reduce cost barriers and promote HET, over-reliance on government funding may strain fiscal resources. To address this, policymakers should diversify funding sources, such as through carbon trading revenues or public–private partnerships, to ensure financial sustainability. At the same time, subsidies should be integrated into existing social welfare frameworks to improve targeting efficiency and reduce administrative costs. Complementary policies, such as stricter energy efficiency standards and public awareness campaigns, can further enhance the effectiveness of subsidies and reduce long-term implementation costs. Moreover, robust monitoring and evaluation mechanisms should be established to adapt policies as economic and technological conditions evolve. By adopting a balanced and adaptive approach, policymakers can maximize the benefits of subsidies while minimizing potential trade-offs.

6. Conclusions

Our paper finds that the LCCP has significantly contributed to HET in China, highlighting the importance of well-designed policies to address climate change and energy shortages through sustainable energy use. As a key initiative aimed at reducing emissions and encouraging cleaner energy adoption, the LCCP provides a valuable case study for understanding the link between policy interventions and household energy transition. While existing research on energy consumption modeling in China primarily focuses on predictive approaches, the connection between policy interventions and household energy transition remains insufficiently explored [48,49,50]. These models provide valuable insights but do not capture how policies like the LCCP influence household energy choices. By incorporating individual and household characteristics, our paper extends previous research by examining how LCCPs shape HET, bridging the gap between predictive modeling and behavioral responses.

Similar energy policies have been implemented to drive energy transition and emission reductions around the world. London’s low-emission zones significantly reduced particulate matter concentrations within regulated areas but had limited effects on nitrogen oxide levels [51]. In the European Union, carbon tax policies have been shown to lower emissions [52]. Emission trading schemes have also contributed to reducing fossil fuel dependence [53]. These international cases highlight the effectiveness of market-based mechanisms and regulatory policies in shaping energy consumption patterns. By comparing China’s LCCP with similar policies, our paper underscores the role of localized policy designs in promoting household energy transition and provides insights for future policymaking beyond China.

However, our paper has limitations. To ensure that the LCCP is effective, government departments have formulated a series of low-carbon development plans and invested numerous financial resources. As our paper confirms, the LCCP does help promote HET and is critical to achieving carbon neutrality. However, the relationship between the initial investment in the policy and its actual effects remains unclear. It is uncertain whether high policy input leads to significant energy conservation and emission reduction effects or if it results in substantial costs with limited effects. Unfortunately, we lack an examination of the costs and benefits of the LCCP and are unable to quantify accurately the benefits of its implementation. This shortcoming should be addressed and improved in future research.