Examining the Role of Renewable Energy, Technological Innovation, and the Insurance Market in Environmental Sustainability in the United States: A Step toward COP26 Targets

Abstract

Investigating the determinants of environmental sustainability has become a very attractive and popular area of research in response to the United Nations’ sustainable development goals (SDGs) and COP26 targets. Against this backdrop, this paper aims to explore the effects of renewable energy, technological innovation, and the insurance market on environmental sustainability in the United States (US). This work contributes to the extant body of knowledge by exploring the effect of the insurance market on the load capacity factor (LC), specifically regarding the US. The LC tracks a certain ecological threshold by simultaneously comparing biocapacity and ecological footprint (EF), thereby providing a comprehensive empirical analysis of ecological sustainability determinants. However, this proxy includes the combined attributes of the demand and supply sides of ecological sustainability. Using the recently developed autoregressive distributed lag method, this research reveals that the insurance market adversely affects the LC in the US. The overall outcomes highlight the positive role of renewable energy, technological innovation, and the insurance market in achieving the SDGs and COP26 targets. Policy recommendations for policymakers concerning promoting renewable energy, green innovation activities, the green insurance market, and ecological regulations are also discussed.

1. Introduction

Ecological economics has focused on environmental problems over the past few decades. Pollution, climate change, and habitat destruction threaten natural biodiversity [1]. According to the World Health Organization (WHO), air pollution is the largest source of pollution in the world, resulting in 1 in 9 deaths worldwide and negatively impacting economies and human well-being [2]. Approximately 7 million people die prematurely every year because of air pollution, including deaths from respiratory infections and lung cancer (WHO 2018). Oceans and ecosystems are threatened by melting glaciers, rising ocean temperatures, and migration to unpolluted waters because of ecological degradation [3]. One of the major threats to the environment is water pollution [4]. The negative effects of water pollution on health and ecosystems lead to economic losses and death because of illnesses caused by drinking contaminated water [5]. As a result of industrialization and urbanization, wastewater is often contaminated, with excessive amounts of toxins appearing in natural substances [6]. It is estimated that 15% of deaths in developed nations are caused by soil pollution, which contains toxic metals and chemicals [7]. The heavy metal contamination of soil threatens human health through contaminated food.

Various international treaties, including the Kyoto Protocol, the Paris Agreement, and the United Nations framework for the convention on climate change, have discussed ways to minimize the harmful effects of air pollution [8]. Researchers have studied the factors influencing carbon dioxide carbon emissions (CO₂) in air pollution [9]. The UN sustainable development goals (SDGs), commonly referred to as the Global Goals, were officially embraced by the United Nations in 2015. They serve as a comprehensive and inclusive appeal to address poverty eradication, ensure environmental preservation, and promote peace and prosperity for all individuals by 2030. These SDGs also require minimizing the problems associated with soil pollution and water, which creates socioeconomic issues [10]. CO₂ emissions are directly related to SDG-13, which concerns climate action, while SDG-14 on below-water life and SDG-15 on land-based life can be achieved by mitigating soil and water pollution, respectively. Ref. [11] proposed using ecological footprint accounting to combine the pollution types in this context. An environmental footprint accounting system comprises the EF and biocapacity. In recent years, the concept of an ecological footprint (EF) has been studied extensively [12,13]. To achieve the SDGs, it would be more accurate to analyze the EF and biocapacity concurrently Ref. [14]. The load capacity factor (LC) proposed by the authors of [15] has been used recently for environmental assessment, with the authors arguing that using the EF alone will not suffice.

According to the Global Footprint Network, biocapacity in the United States (US) is 3.4 gha and the EF is 8.0 gha. Hence, the US exhibits the highest ecological shortage. For example, in 2021, petroleum sources provided approximately 90 percent of the country’s total energy use in the transportation sectors [16]. Figure 1 shows the types and amounts of primary energy sources consumed in the US. The country obtains 77% of its total energy from fossil fuel utilization (oil, natural gas, and coal). The US has also increased its C${\mathrm{O}}_2$ emissions and ecological footprint over the years. Figure 2 shows the variations in ecological footprint in the US from 1983 to 2020. It shows that the ecological footprint increased from 8.88 gha per capita in 1983 to 9.51 gha per capita in 2009. Conversely, the ecological footprint in the country decreased to 7.06 gha per capita in 2020. The country still has one of the highest ecological footprints in the world.

The US produced 14% of the world’s CO₂ in 2020. There is evidently a serious environmental problem in the US and reducing pollution should be a priority. Therefore, considering the country’s rapid economic activity and its contribution to environmental pollution, it is essential to investigate the factors affecting environmental pollution in the US. Using econometric analysis, this work examines the available data from 1983 to 2020. Based on biocapacity/EF, the LC is the inclusive measure of climatic quality. The field of LC literature is an emerging one, and researchers are using this factor in different data analysis techniques. This work aims to probe the effects of technological innovations, the insurance market, renewable energy consumption (REC), and income on the LC in the US. Using the recently developed autoregressive distributed lag method (ARDL), this research shows that REC, the insurance market, and technological innovation promotes ecological quality in the US, whereas economic growth adversely affects ecological quality in the US. The overall outcomes of this study highlight the positive role of the insurance market, REC, and technological innovation in achieving the SDGs and the climate change conference (COP26) targets.

Today, the world is heading towards achieving the SDGs through technological innovations and renewable energy usage. COP26 requires that governments invest more in clean energy sources, providing maximum employment opportunities. The primary objective of COP26 was to facilitate consensus among all signatories of the Paris Agreement regarding the formulation and presentation of their respective nationally determined contributions (NDCs), which were aimed at mitigating greenhouse gas emissions. The insurance market plays a crucial part in providing individuals and markets with a sense of security and protection. The provision of indemnity by the insurance industry significantly impacts the acquisition and utilization of various goods, such as electronic equipment, motor vehicles, and homes [17]. Individuals purchase energy-consuming goods to enhance their quality of life, relying on the assurance of a risk-free income through insurance. However, it is important to acknowledge that the utilization of such goods can have detrimental effects on the climate [18]. Consequently, insurance market has the potential to exert an influence on consumption, investment, and climate-changing decisions made by the people. According to a previous study [19], the insurance industry plays a crucial role in strengthening investors’ investments, thereby contributing to the advancement of financial positions in markets. Therefore, insurance activities have the potential to contribute to higher levels of energy consumption, thereby exerting a detrimental impact on environmental quality. Ref. [20] proposes the establishment of a connection between insurance, energy, and the environment through the allocation of increased income and profits by businesses and households toward the sale, purchase, and utilization of energy-intensive commodities. This arrangement would be accompanied by risk mitigation measures that are offered by the insurance sector. However, the development of the insurance sector can be useful to the environment. In this context, the insurance market can play a positive role in promoting clean energy sources and by facilitating environmental technology development and can also reduce costs by creating a stable and risk-free investment environment.

Energy, gross domestic product (GDP), REC, and carbon emissions have been controversial topics in environmental research over the last few decades. The significant impact of green energy on ecological sustainability can be attributed to the reduction of CO₂ in the atmosphere. The drivers of ecological sustainability have been studied empirically as a result. According to the authors of [16], the GDP of sub-Saharan Africa has a negative effect on the climate. The authors demonstrated a positive correlation between GDP and CO₂ in Africa. The authors of [21] suggested that CO₂ affects the GDP of China. The authors of [22] have also shown that income changes increase CO₂ in Finland. Elsewhere, Ref. [23] illustrated that income is related to CO₂ in Chile. An increase in revenue in China increases CO₂ dioxide production, according to the authors of [24]. In Turkey, the authors of [25] found that a rise in GDP improved the environment. Another work by [26] concluded that in the G7 countries, income increased the levels of CO₂. Additionally, the authors of [27] showed that GDP positively impacts CO₂, while [28] showed that income enhances EF in the Organization for Economic Co-operation and Development (OECD) countries. The same results were reported in other studies [29,30]. Ref. [31] studied the connection between REC and CO₂ in 74 economies. The authors discovered a positive connection between REC and CO₂. Finally, Ref. [32] indicated that REC promotes ecological sustainability.

Ref. [33] concluded that REC significantly and positively impacted Turkish ecology. In 15 Asian economies, REC also positively affected ecological sustainability. Ref. [34] concluded that the REC degrades the climate in the case of China. Ref. [35] proposed that an upsurge in RE will lead to sustainability in 187 countries. Many research studies have examined the relationship between GDP, REC, CO₂, and EF, but few have assessed how GDP and REC affect LC. This factor is used as a comprehensive and complete measure of the environment. Ref. [14] evaluated the impact of REC on the LC in the US. The authors showed that REC impacts the LC positively. Moreover, the authors of [36] stated that RE reduces the LC, whereas GDP is positively correlated. Ref. [37] concluded that RE positively affects the LC in South Africa, while GDP negatively affects the LC in South Korea, and G7 countries are non-linearly related, according to the authors of [38].

Very scant studies have calculated the effect of insurance on CO₂. Ref. [19] estimated the impacts of the insurance market on the levels of environmental quality in OECD. Based on panel data, the authors of [39] showed that insurance expansion raised CO₂ in the nations of Brazil, Russia, India, China, and South Africa (BRICS). Using a nonlinear ARDL, the authors of [40] found that the insurance market and REC lower pollution in Russia, China, and South Africa. Conversely, the researchers showed that Russian insurance increases CO₂ levels.

Only a few research groups have assessed the interrelationship between the insurance industry and the environment. These studies argue that these markets can improve climatic quality. The riskless income that is financed by insurance can use environmental resources to improve clean energy innovations and minimize environmental degradation. Nevertheless, the authors of [18] found that insurance markets have contributed to environmental pollution. Hence, there is disagreement regarding the impacts of the insurance sector on the climate. Furthermore, a paper has yet to examine the effect of insurance on the LC.

To address this knowledge gap, the present paper examines the effects of the insurance market on the LC in the US. This empirical work contributes to the literature by providing new insight into the relationship between the LC and the insurance market in the case of the US by addressing the following questions: (1) How do the insurance markets impact LC? and (2) Does REC enhance LC?

The structure of our work is organized as follows. The second section shows the materials and methods used in the study, while the third and fourth sections show the empirical findings and offer our conclusions.

2. Materials and Methods

2.1. Modeling and Data

We used the annual data recorded from 1983 to 2020 to evaluate the effect of technological innovation, renewable energy, the insurance market, and economic growth on the load capacity factor in the US. To this end, we performed an empirical analysis by focusing on the LC as a novel proxy to capture ecological sustainability, and the empirical model is formulated as follows:

$${\mathrm{lnLC}}_{,\mathrm{t}}={\vartheta }_0+{\vartheta }_1{\mathrm{lnGDP}}_{\mathrm{t}}+{\vartheta }_2{\mathrm{lnREC}}_{\mathrm{t}}+{\vartheta }_3\mathrm{lnTI}+{\vartheta }_4{\mathrm{lnIM}}_{\mathrm{t}}+{\mathrm{e}}_{\mathrm{t}}$$

(1)

where ${\mathrm{LC}}_{\mathrm{t}}$ is the load capacity factor, ${\mathrm{lnGDP}}_{\mathrm{t}}$ represents economic growth, ${\mathrm{lnREC}}_{\mathrm{t}}$ represents renewable energy consumption, and ${\mathrm{lnIM}}_{\mathrm{t}}$ represents the insurance market. A description of the focused variables and data sources is presented in

Table 1. Description of the focused variables and data sources.

Variable	Variable Description	Source
$\mathrm{LC}$	This indicator is captured by dividing the biocapacity and EF per capita (global-hectares)	Global Footprint Network (2023)
$\mathrm{IM}$	The total insurance $\left(\mathrm{life} \mathrm{and} \mathrm{non}-\mathrm{life}\right)$ as a percentage of GDP	OECD (2023)
TI	Total patents (addition of resident and nonresident patents)	World Bank (2023)
$\mathrm{REC}$	Renewable energy per capita	Our World in Data (2023)
$\mathrm{GDP}$	GDP per capita (Constant-2015 United States-dollars)	World Bank (2023)

.

2.2. Methods of Estimations

The classical tests of unit roots, such as the augmented Dickey–Fuller assessment, ignore any structural changes (SC). Hence, this paper uses the assessments developed by Perron and Vogelsang (1992) [41] and Zivot and Andrew (1994) [42], with one SC. The paper also uses the ARDL approach to assess the long-term association between GDP, REC, TI, IM, and LC. This method was suggested in [43]. The main advantage of this test is that it is more suitable for small amounts of tested data. However, the Akaike information criterion (AIC) is used to select the lag length (p). In addition, this method evaluates cointegration via three options: I(0), I(1), or both I(0) and I(1). In this method, if the ${\mathrm{F}}_{\mathrm{Statistics}}$ value exceeds the upper bounds of I(0) and I(1), the alternative hypothesis, ${\mathrm{H}}_1$, will be statistically accepted. However, if the $\mathrm{F}$ statistics value falls between I(0) and I(1), this indicates that the cointegration results are inconclusive.

The main disadvantage of the ARDL bound test, as suggested by Pesaran (2001) [43], is instability issues regarding the outcomes. This instability issue can be fixed using the bootstrapping ARDL method approach suggested by McNown et al. (2018) [44], thereby increasing the power of the “t-test” and “F-test.”

The classical ARDL only shows the critical values (CVs) for F-tests and the value of t for dependent tests, ignoring the “F” independent tests, while the upgraded versions propose an additional test on the lagged independent examined variable(s) to complement the current (F) and (t) tests of the ARDL approach, as advanced by Pesaran et al. (2001) [39]. This advantage will help to overcome the instability problem in the ARDL findings. In this context, the CV of the classical ARDL allows for the endogeneity of one explored variable. In contrast, the CV generated in the upgraded ARDL method, as advanced by McNown et al. in 2018, combines the endogeneity of specified explanatory variables. This technique is also suitable for the tested models that include more than one explanatory tested variable. The co-integration amid the focused variables will be captured if the values of $\mathrm{F}{-}_{\mathrm{Pesaran}}$, $\mathrm{t}{-}_{\mathrm{dependent}}, \mathrm{F}{-}_{\mathrm{independent}}$ exceed the CVs achieved using the approach developed by McNown et al. in 2018. Sam et al. (2019) [45] proposed the augmented ARDL model (AARDL) as a new extension of the bootstrap ARDL approach. According to the AARDL approach, the CV of the “$\mathrm{F}$” statistics for the examined samples can be generated from Narayan (2005) [46]. Likewise, the “t” statistics values can be obtained from Pesaran et al. (2001) [43]. Thus, the three hypotheses are as follows.

First, the null hypothesis (${\mathrm{H}}_0$) is used for the overall F test on all variables: ${\mathrm{H}}_0$: ${\mathsf{\sigma }}_1$ = ${\mathsf{\sigma }}_2$ = ${\mathsf{\sigma }}_3$ = ${\mathsf{\sigma }}_4$ = ${\mathsf{\sigma }}_5$ = 0, while the alternative hypothesis $\left({\mathrm{H}}_1\right)$: ${ \mathsf{\sigma }}_1\ne { \mathsf{\sigma }}_2$ ≠ ${\mathsf{\sigma }}_3 \ne {\mathsf{\sigma }}_4\ne { \mathsf{\sigma }}_5 \ne$ 0.

Second, the H0 for the t-test on only the dependent variable is: ${\mathsf{\sigma }}_1$ = 0, while ${\mathrm{H}}_1$: ${\mathsf{\sigma }}_1$ ≠ 0

Third, the ${\mathrm{H}}_0$ of the F-test on independent variables is:${ \mathrm{H}}_0$: ${ \mathsf{\sigma }}_2$ = ${\mathsf{\sigma }}_3$ = ${\mathsf{\sigma }}_4$ = ${\mathsf{\sigma }}_5$ = 0, while ${\mathrm{H}}_1$: ${\mathsf{\sigma }}_2$ ≠ ${\mathsf{\sigma }}_3 \ne {\mathsf{\sigma }}_4\ne { \mathsf{\sigma }}_5 \ne$ 0.

The upgraded method is formulated in the next equation:

$$\begin{gathered} \Delta {\mathrm{lnLC}}_{\mathrm{t}}={\mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_1\Delta {\mathrm{lnLC} }_{2\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_2\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_3\Delta {\mathrm{lnREC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_4\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{y}}_5\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega }\mathrm{ECT}}_{\mathrm{t}-1 }+{\mathsf{\epsilon }}_{1\mathrm{t}}. \end{gathered}$$

(2)

In the above equation, ∆ represents the first difference operator, ${\mathrm{lnL}}_{\mathrm{t}}$, ${\mathrm{lnGDP}}_{\mathrm{t}}$, ${\mathrm{lnREC}}_{\mathrm{t}}$, ${\mathrm{lnTI}}_{\mathrm{t}},{ \mathrm{and} \mathrm{lnIM}}_{\mathrm{t}} \mathrm{are}$ the focused variables, n represents the optimal lags, and ${\mathsf{\epsilon }}_{\mathrm{t}}$ refers to the error term. The error-correction model $\left(\mathrm{ECM}\right)$ is employed to assess the velocity of adjustment among the short- and long-period levels of the dependent variable. In the equation, ${\mathrm{ECT}}_{\mathrm{t}-1}$ is the one period of lagged ECT, which should be statistically significant and shown with (−) as a negative sign.

Likewise, we utilized the Ramsey-reset, LM, Breusch–Pagan–Godfrey, ARCH, and normality assessments to ensure that the employed model is statically correct and stable. To evaluate the causality of the selected variables among $\mathrm{L}\mathrm{C}$, GDP, REC, TI, and IM, Granger causality analysis was used. This analysis aims to identify if there is a causal connection. In this test, ($\mathrm{ECT}$) is employed to evaluate the short-term deviations of the employed series from the long-term equilibrium path. The ECM equation is calculated as follows:

$$\begin{gathered} \Delta {\mathrm{lnLC}}_{\mathrm{t}}={ \mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_1\Delta {\mathrm{lnLC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnREC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega } \mathrm{ECT}}_{\mathrm{t}-1}+{\mathsf{\epsilon }}_{1\mathrm{t}} \end{gathered}$$

(3)

$$\begin{gathered} \Delta {\mathrm{lnGDP}}_{\mathrm{t}}={ \mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_1\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnLC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnREC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega } \mathrm{ECT}}_{\mathrm{t}-1}+{\mathsf{\epsilon }}_{1\mathrm{t}} \end{gathered}$$

(4)

$$\begin{gathered} \Delta {\mathrm{lnREC}}_{\mathrm{t}}={ \mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_1\Delta {\mathrm{lnREC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnLC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega } \mathrm{ECT}}_{\mathrm{t}-1}+{\mathsf{\epsilon }}_{1\mathrm{t}} \end{gathered}$$

(5)

$$\begin{gathered} \Delta {\mathrm{lnTI}}_{\mathrm{t}}={ \mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_1\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnLC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnRECI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega } \mathrm{ECT}}_{\mathrm{t}-1}+{\mathsf{\epsilon }}_{1\mathrm{t}} \end{gathered}$$

(6)

$$\begin{gathered} \Delta {\mathrm{lnIM}}_{\mathrm{t}}={ \mathsf{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_1\Delta {\mathrm{lnIM}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnLC}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnGDP}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnRECI}}_{\mathrm{t}-\mathrm{j}}+ \\ {\sum }_{\mathrm{i}=1}^{\mathrm{n}}{}^{\prime }{\rm Y}_0\Delta {\mathrm{lnTI}}_{\mathrm{t}-\mathrm{j}}+{\mathsf{\omega } \mathrm{ECT}}_{\mathrm{t}-1}+{\mathsf{\epsilon }}_{1\mathrm{t}}. \end{gathered}$$

(7)

In Equations (3)–(7), ∆ represents the operator of the first difference, $\mathrm{lnLC}, \mathrm{lnGDP}, \mathrm{lnREC}, \mathrm{lnTI}, \mathrm{and} \mathrm{lnIM},$ are the tested variables of this work, and ${\mathsf{\epsilon }}_{\mathrm{t}}$ represents the error term. Figure 3 shows the structure of the employed methods.

3. Results and Discussion

3.1. Empirical Results

The descriptive statistics of the focused variables that were induced (the mean, median, maximum, minimum, and standard deviation) are presented in

Table 2. Descriptive statistic results.

	LC	GDP	REC	TI	IM
Mean	−0.825898	10.75480	1.528109	316030.8	13.96549
Median	−0.809167	10.79795	1.461889	280829.0	14.06364
Maximum	−0.634108	11.01367	2.042558	606956.0	14.89216
Minimum	−0.971521	10.37015	1.133224	58457.76	12.43639
Std. Dev.	0.092004	0.181216	0.229573	177414.9	0.719693

. In the first step of the empirical process, we employed the Zivot and Andrew (ZA) and Perron and Vogelsang (BA) assessments with structural changes in the unit root tests to discover the integration order of the selected variables for use in the cointegration assessments.

The results from these tests, as presented in

Table 3. Findings of the Perron and Vogelsang and Zivot and Andrew assessments.

P.V. Assessment	Level		${\mathbf{First}}_{\mathbf{difference}}\Delta$.
Variables	$Tes{t}_{Statistics}$	SC		$Tes{t}_{Statistics}$	SC
lnLC	−2.476986	2011	$\mathsf{\Delta }$lnLC	−6.399919 ***	2008
lnGDP	−3.901480	1996	$\mathsf{\Delta }$lnGDP	−6.122386 ***	2008
lnREC	−2.684303	2010	$\mathsf{\Delta }$lnREC	−7.757725 ***	2001
lnTI	−2.087935	2018	$\mathsf{\Delta }$lnTI	−9.564790 ***	2019
lnIM	−1.965517	1996	$\mathsf{\Delta }$lnIM	−13.43329 ***	1988
ZA Test	Level		$Firs{t}_{difference}\Delta$
Variables	$Tes{t}_{Statistics}$	SC		$Tes{t}_{Statistics}$	SC
lnLC	−2.949537	1993	$\mathsf{\Delta }$lnLC	−5.968805 ***	1993
lnGDP	−2.619507	1996	$\mathsf{\Delta }$lnGDP	−5.318506 ***	2007
lnREC	−3.487440	1998	$\mathsf{\Delta }$lnREC	−6.780444 ***	1991
lnTI	−1.903031	2014	$\mathsf{\Delta }$lnTI	−6.392394 ***	1998
lnIM	−2.568962	2009	$\mathsf{\Delta }$lnIM	−5.483452 ***	1997

*** means the significance of variables at 1 percent. SC: structural changes.

, show that all the examined series are not stationary in their level values. It can be seen from Table 3 that all series became stationary at the first operation level. Based on these outcomes, it can be concluded that the ARDL approach can be used since the variables used in this work are stationary. Figure 4 shows the tested variables in the plots.

The empirical outcomes of the AARDL and diagnostic tests are displayed in

Table 4. The augmented ARDL approach.

Test ‘Statistics’
${\mathrm{F}}^{\mathrm{overall}}$			${\mathrm{T}}^{\mathrm{dependent}}$			${\mathrm{F}}^{\mathrm{independent}}$
6.904689			−4.712924			7.347993
critical values	‘1%’		‘5%’		‘10%’
$\mathrm{Statistics}$	${\mathrm{I}}_{\left(0\right)}$	${\mathrm{I}}_{\left(1\right)}$	${\mathrm{I}}_{\left(0\right)}$	${\mathrm{I}}_{\left(1\right)}$	${\mathrm{I}}_{\left(0\right)}$	${\mathrm{I}}_{\left(1\right)}$	${\mathrm{Test}}_{\mathrm{Reference}}$
${\mathrm{F}}^{\mathrm{overall}}$	3.$74$	5.$06$	2.86	4.$01$	2.$45$	3.$52$	[46]
${\mathrm{T}}^{\mathrm{dependent}}$	−3.$43$	−4.$60$	−2.$86$	−3.99	−2.$57$	−3.$66$	[43]
${\mathrm{F}}^{\mathrm{independent}}$	3.$33$	5.$47$	2.$39$	4.$18$	1.$96$	3.$58$	[45]
Diagnostic
Tests	${\mathrm{F}}_{\mathrm{statistics}}$	$p_{\mathrm{value}}$			Tests	${\mathrm{F}}_{\mathrm{statistics}}$	$p_{\mathrm{value}}$
Rams${\mathrm{ey}}^{\mathrm{test}}$	1.444687	0.1667			ARC${\mathrm{H}}^{\mathrm{test}}$	0.157629	0.6942
Normalit${\mathrm{y}}^{\mathrm{test}}$’	1.873994	0.3918			Heteroskedasticit${\mathrm{y}}^{\mathrm{test}}$	1.344302	0.2737

. According to the findings of this test, the values of the “F” overall statistics, “F” independent, and “t” dependent tests implied a rejection of the hypothesis of no cointegration among (GDP, REC, TI, IM, and LC). This means that these results affirm the presence of cointegration among the selected variables. Furthermore, the J-B test affirmed the normality of the employed model’s residual terms. In addition, the Ramsey test, ARCH, and heteroskedasticity tests confirmed that the employed model of this work is free from autocorrelation and is significantly stable and correct. Likewise, the CUSUM and CUSUMsq (Figure 5) analysis implies the stability of the long-term parameters.

Table 5. ARDL short- and long-term outcomes.

Variable	Coefficient	$\mathbf{Std}. \mathbf{Error}$	${\mathbf{T}}_{\mathbf{Statistics}}$.	Prob.
Short term
$\mathsf{\Delta }$lnGDP	−1.514703 ***	0.421614	−3.592632	0.0021
$\mathsf{\Delta }$lnREC	0.263253 ***	0.061561	4.276310	0.0005
$\mathsf{\Delta }$lnTI	0.102612	0.090170	1.137992	0.2700
$\mathsf{\Delta }$lnIM	0.359209 **	0.130039	2.762324	0.0128
Long term
lnGDP	−2.128675 ***	0.389068	−5.471218	0.0000
lnREC	0.369961 ***	0.032510	11.37996	0.0000
lnTI	0.144205 **	0.054742	2.634267	0.0168
lnIM	0.504811 ***	0.105952	4.764546	0.0002
${\mathrm{ECT}}_{\mathrm{t}-1}$	−0.711571 ***	0.107131	−6.642070	0.0000

Note: ** and *** indicates the significance of the study variables at 5%, and 1% levels.

displays the empirical findings using the ARDL method. We found that economic growth declines the LC significantly. At the 1% significance level, a one-percent increase in GDP decreases the LC by 2.128675% and 1.514703% in the long and short terms. These obtained outcomes illustrate the finding that GDP negatively affects environmental quality. Moreover, the findings demonstrate that the coefficient of technological innovation is positive and significant in the long term; however, this coefficient is not significant over a short period. The outcomes affirm that technological innovation positively affected environmental neutrality in the long term.

Furthermore, we found that REC promotes the LC significantly. At the 1% significance level, a one-percent increase in REC promotes the LC by 0.369961% and 0.263253% in both the long term and the short term. These findings affirm that REC positively affects the ecological neutrality in the country for the focused data.

The findings from ARDL also show that the insurance market promotes LC in the short term. A one-percent increase in the insurance market increases the LC by 0.504811% and 0.359209% in the long and short term. Thus, the findings affirm that the insurance market positively impacts environmental neutrality in the US. Figure 6 shows a summary of the findings when employing the ARDL approach.

The coefficient of ECM 1 is statistically negative (−0.9633), which corroborates the long-term link among GDP, REC, TI, IM, and LC.

Table 6. Granger causality outcomes.

Causality	${\mathbf{F}}_{\mathbf{Statistic}}$	p-Value
GDP→LC	7.71175	0.0089
LC→GDP	0.02474	0.8759
REC→LC	6.62543	0.0146
LC→REC	0.14208	0.7086
TI→LC	0.35140	0.5572
LC→TI	2.22689	0.1448
IM→LC	6.78676	0.0135
LC→IM	1.13586	0.2940

presents the Granger causality analysis. The findings of this test show evidence of unidirectional causality from IM to LC, and also show unidirectional causality from GDP to LC, and from TI to LC. However, these findings reveal a significant link among GDP, TI, IM, and LC.

3.2. Discussion of Findings

Many meetings and protocols have been organized to address ecological sustainability, such as the Kyoto Protocol (1997), the Paris Agreement (2015), and, more recently, the COP-26 conference (2021); these meetings and agreements aim to combat ecological pollution and cut global warming to well below ‘2 °C’ above pre-industrial levels. In addition, many scholars have evaluated the determinants of environmental sustainability. However, ecological indicators such as carbon emissions and EF cannot fully reflect ecological sustainability. In this regard, this study focuses on the load capacity factor (LC). However, the study aims to evaluate technological innovation, REC, the insurance market, and economic growth in environmental sustainability in the US. To this end, the study takes the data from 1983–2020 to explore the impact of technological innovation, REC, the insurance market, and economic growth on environmental sustainability in the US. To achieve this, our study employed the Perron and Vogelsang (1992) [37] and the Zivot and Andrew (1994) [38] unit root tests to determine the integration order among the examined variables. In addition, a novel approach of augmented ARDL analysis is employed to evaluate the cointegration present in the employed model. Then, the ARDL approach was utilized to evaluate the short- and long-term links among the selected variables.

Moreover, Granger causality analysis was utilized to explore the causality connections between the variables. This work contributes to the extant body of knowledge by exploring the impact of the insurance market on the LC, specifically regarding the US. The LC tracks a certain ecological threshold by comparing “biocapacity” and “EF” simultaneously.

The results determine that REC significantly promotes the LC in both the short term and long term. Hence, REC plays a crucial role in achieving COP-26 targets. These findings are supported by the findings of [19]. The authors evaluated the impact of RE on the LC in the OECD and showed that RE impacts the LC positively.

Likewise, the outcomes show that economic growth lessens environmental quality in both the short term and long term; these outcomes align with the findings reported by [17], who assessed the impact of economic growth on the LC in the US. The obtained outcomes imply that higher income does not inevitably result in more environmental neutrality. Thus, there is a need for the US to adopt energy-related environmental policies. Although the US has the potential to lead a global transition to REC, the energy sector in this country is heavily dependent on fossil fuels. The US obtains 78% of its total energy from fossil fuel sources. Hence, the study suggests that the policymakers in the US should increase their investment in renewable energy sources such as solar and wind energy. This investment in renewable energy sources will support environmental quality and achieve COP-26 targets.

Besides, the outcomes show that technological innovation and the insurance market have a positive impact on LC in both the short term and the long term. These outcomes align with the findings reported in previous studies [47,48,49]. These studies suggested that environmental innovation contributes to environmental sustainability. However, this study’s findings show that improving technological innovation over time will lead to better technologies that require fewer production resources, which, in turn, mitigates ecological degradation. Therefore, our findings affirm the positive role of technological innovation and COP26 targets. These outcomes are reasonable since technological innovation might provide novel and efficient clean technologies, enhancing productivity in the economy using limited resources. However, the availability of green technologies can also help to promote green energy initiatives.

4. Conclusions

In recent years, many meetings and agreements have been set up to address global warming challenges, such as the United Nations (UN) conference in 1972, the Kyoto Protocol in 1997, the Paris Agreement in 2015, and the COP-26 2021 conference in the United Kingdom. The COP-26 conference urged further commitment to the Paris Agreement, particularly by governments worldwide. This conference was the 26th meeting of the UN Framework Convention on Climate Change, established to evaluate the efforts and policies made by various countries to address the challenge of global warming. One of the most significant breakthroughs in the conference was the resolution to affirm sustained funding for progression toward the Paris 2015 Agreement and the United Nations framework resolution on global warming, mitigating the global temperature by up to 2 °C. Keeping in mind the importance of the US economy as the primary economic player and a significant contributor to global warming, it is indispensable to identify effective policies to mitigate ecological pollution in the US, which is committed to making the COP26 conference a turning point of global efforts to address these climate challenges. To this end, we took the data from 1983 to 2020 to explore the impacts of technological innovation, REC, the insurance market, and economic expansion on ecological sustainability in the US. Our study contributes to the extant body of knowledge by exploring the impact of the insurance market on LC, specifically regarding the US.

As discussed in Section 3, we used the recently developed ARDL to illustrate the point that economic growth negatively affects the LC in both the long and short terms. Therefore, economic expansion has an adverse effect on the US’s ecological quality. Likewise, the findings prove that REC, technological innovations, and the insurance market have positive effects on the LC. These outcomes confirm the positive role of REC, the insurance market, and technological innovations in achieving the SDG and COP-26 targets. On the other hand, the study’s outcomes affirm that economic growth has a negative effect on achieving the SDG and COP-26 targets. These outcomes could be explained by the fact that renewable energy resources, technological innovation, and the insurance market are essential for the sustainable development of the US economy. However, the insurance market can play a positive role in promoting environmental quality by increasing capital access and funding renewable energy projects. In addition, the insurance market can play a positive role in the promotion of green technological development, which, in turn, will decrease the utilization of fossil fuel sources and promote environmental quality. Hence, to achieve the COP-26 targets, policymakers must use the insurance market’s growth and technological innovations to promote ecological sustainability.

Based on the presented analysis, this study demonstrates that the variables of RE, TI, and IM will improve the climate of the USA, while the GDP degrades it. The following are a few recommendations.

• Greener growth strategies should be pursued by the US government. Technology innovation should be leveraged in environmentally friendly sectors to increase the LC. Policies that can contribute to green growth in the US economy include carbon capture technologies, the cost reduction of renewable energy technologies, more high-tech and efficient solar panels, and wind turbine incentives. In the same way, the US government can limit the decline in the LC by imposing environmental taxes on polluting companies and groups. There is evidence that a significant improvement in the IM promotes new investment activities that will adversely affect the environment by promoting fossil fuel utilization.
• To promote LC growth, the government should enhance its investment, innovation, and infrastructure spending on renewable energy. Policymakers should increase the share of renewables in terms of total energy. SDG-7—regarding clean and affordable energy—is associated with renewable resources; therefore, SDG-7 can strengthen the LC. Governments need to make renewable resources more cost-effective in this context. Moreover, the government should promote RE in households and industries through environmental awareness programs. With the use of renewables, the US can improve environmental quality, as well as create jobs and increase energy security.
• Third, the policymakers must use the development of the insurance sector to promote the environment. Therefore, clean energy sources must be effectively financed by the insurance industry in the US. Additionally, insurers should facilitate environmental technology development and reduce costs by creating a stable and risk-free investment environment. Insurers may harm their own interests if they act without regard for the environment since having a poorer environment has a negative impact on society’s health and economy. As a result, governments must take action to expand green insurance financing options. In the US, the current insurance activities reduce the LC. Non-green insurance investments and activities need to be limited by governments. In addition, financial incentives should be provided for sustainable green insurance investments.

In conclusion, we contribute to the extant body of knowledge by exploring the impact of the insurance market on the load capacity factor, specifically regarding the US. The primary limitation of this work is that the data employed cover the period from 1983 to 2020. We had hoped to extend our tested data coverage to the most recent year, 2022, but data constraints have forced us to confine the data employed to 2020. Likewise, we have used the AARDL approach to evaluate the links among the focused variables. Future studies can employ other methods, such as NARDL, to examine the impact of the insurance market on ecological quality.