Impact of Artificial Intelligence,Health Expenditure and Digital Financial Inclusion on Life Expectancy

Abstract

The Chinese government has actively promoted artificial intelligence (AI) in health care, with momentum building in 2016 through the Healthy China 2030 Initiative. This reform plan aims to modernize health care infrastructure, reduce the burden of chronic diseases, and expand access to medical services in rural areas using digital technologies such as AI. Health expenditure (HE) and digital financial inclusion (DFI) play a crucial role in improving health outcomes. HE improves access to medical services and the quality of care, while DFI allows individuals to afford health care, save for emergencies, and manage health-related financial risks. Therefore, this study examines the impact of AI, health expenditure and DFI on life expectancy (LE) in China from 2013Q1 to 2023Q4. This study employed autoregressive distributed lag (ARDL) and quantile regression analyses to ensure the robustness of the results. The finding shows that gross domestic product (GDP), AI, health expenditure, DFI and government effectiveness have positive effect on LE. This study recommended that the government expand AI integration in health care to improve diagnostics and treatment efficiency. It also emphasized promoting DFI to help low-income groups access health care and manage medical expenses. As well, the study suggested increasing public health expenditure to enhance health care infrastructure and service quality, ultimately improving LE.

Keywords

artificial intelligence health expenditure financial inclusion life expectancy ARDL China

Introduction

Living a long and healthy life is a primary goal for both individuals and governments across all populations (Ronmi et al., 2023). A longer lifespan serves as an important indicator of a nation’s overall health, well-being, and socioeconomic development and inclusive growth (A. F. Khan et al., 2025; Uddin et al., 2024). The term life expectancy (LE) denotes the number of years an individual is expected to live. It is defined as the estimated average age at which members of a specific population group are likely to die (Ortiz-Ospina, 2017). LE at a given age represents the average additional number of years a person of that age is expected to live, assuming they experience the current mortality conditions (age-specific death rates) throughout their remaining lifetime. When expressed from age zero, LE at birth refers to the average number of years a newborn is expected to live if subjected to the prevailing mortality rates at each age (Eurostat, 2025). Figure 1, shows life expectancy at birth in China from 1960 to 2023. From 1960 to the mid-1970s, life expectancy rose sharply from 33.42 years in 1960 to around 60.95 years in 1975 reflecting major public health improvements, better nutrition, and reduced infectious diseases. From the late 1970s to the early 2000s, growth continued steadily, reaching around 72 to 74 years by 2005–2006, due to ongoing medical advancements and socioeconomic development. From 2010 onward, life expectancy increased gradually, peaking at 78.20 years in 2022, with a slight decline to 77.95 years in 2023, likely influenced by health challenges such as the COVID-19 pandemic.

Figure 1.

Trends of life expectancy at birth, total (years).

The automation of digital systems, particularly in health care, plays a crucial role in transforming patient care services. Given the complexity of health care systems and the challenges in achieving efficient processes, digital transformation is essential for innovation and improvement. Artificial intelligence (AI) is one of the most disruptive and powerful innovations in modern computer science, with the potential to influence various sectors and change existing practices. In health care, AI continues to attract attention from scholars and policymakers due to its impact on automation. Consequently, understanding AI’s role in automating digital health system technologies is essential (Mohammed et al., 2022). AI is a field of computer science focused on creating machines that can perform tasks typically requiring human intelligence. By learning from data (machine learning) or deriving insights from it (data science), AI achieves human-like decision-making and is applied across various disciplines, including information systems (Ronmi et al., 2023). AI is set to transform medicine, offering the potential to enhance outcomes and experiences for both clinicians and patients (Rajpurkar et al., 2022). The potential of AI in health care has received considerable attention, with applications spanning numerous medical domains. This interest is particularly relevant as health care systems worldwide seek to achieve the ‘quadruple aim’: enhancing patient care experiences, improving population health, reducing per capita health care costs, and supporting the well-being of health care providers (Kelly et al., 2019). AI in China’s health care sector gained momentum in 2016 with the Healthy China 2030 Initiative, a comprehensive reform plan aimed at modernizing health care infrastructure, reducing the burden of chronic diseases, and expanding access to care in rural areas through digital technologies like AI. The National Strategy for AI Development, launched in 2017, further aimed to position China as a global leader in AI by 2030, accelerating applications in medical imaging, diagnostics, and treatment planning (Daxue Consulting, 2025). Alongside these advancements, China introduced regulations to safeguard data security and privacy. The Personal Information Protection Law (PIPL), effective in 2021, established strict rules on the use of personal health data, balancing patient privacy with the development of data-driven AI technologies. While domestic biodata cannot be used abroad, China can access publicly available global biodata, enabling the country to develop proprietary medical applications and devices based on its own collected data sets (Daxue Consulting, 2025).

Public health spending refers to the portion of total health expenditures funded by the government. Increases in public health expenditures improve access to health services, allowing individuals to meet their health care needs more easily (Beylik et al., 2022). Public health expenditure generally includes both health and health-related spending, defined by their primary purpose of improving health, regardless of the main function of the entity providing or financing the services. It encompasses the health of individuals as well as groups or populations. Health expenditure covers all spending on medical care, prevention, health promotion, rehabilitation, community health activities, health administration and regulation, and capital investments aimed primarily at improving health (Oniore et al., 2024). The United Nations Capital Development Fund (UNCDF) suggests that access to financial accounts allows individuals to obtain credit more easily, expand and sustain their businesses, invest in education and health, and manage financial shocks, thereby enhancing livelihood sustainability. Financial inclusion also directly affects people’s health. By providing a means to cover unexpected medical expenses through savings, it can improve health outcomes. In addition, financial inclusion enables access to higher-quality health inputs, such as nutritious food, clean energy, and better sanitation. It also reduces mental stress by offering financial stability, which contributes to overall health and well-being (Xiao & Tao, 2022). Digital financial inclusion (DFI) can play a crucial role in reducing the economic and social impacts of the COVID-19 pandemic. Expanding financial access for low-income households and small businesses may support a more inclusive financial recovery. However, these benefits are not guaranteed, as the pandemic can worsen existing financial exclusion and create new risks in the use of digital financial services. Therefore, implementing mechanisms that promote DFI in both developed and developing countries is essential to advance financial inclusion at various stages of development and make significant progress toward achieving the Sustainable Development Goals (Tay et al., 2022). Financial inclusion (FI) plays a critical role in improving population health outcomes, as financial exclusion limits individuals’ ability to respond to health emergencies such as illness. By increasing income levels, FI enables better access to health care. It also promotes savings for health-related emergencies, contributing to improved health outcomes. Moreover, FI allows individuals to share risk during income or health shocks by providing access to resources such as remittances and credit (Acheampong & Tetteh, 2024).

The main objective of this study is to examine the impact of artificial intelligence, health expenditure and digital financial inclusion on life expectancy in China. By focusing on these three dimensions together, the study aims to provide a deeper understanding of how digital transformation and health investment shape population health in China. This study contributes to the existing literature in several important ways. First, it develops a more comprehensive measure of artificial intelligence. Earlier work relied mainly on AI investment alone, which provides only a partial picture. In contrast, this study constructs an AI index using AI investment, AI patents and AI publications, and applies PCA to generate a stronger and more representative indicator. Second, while previous research on AI and health outcomes has centered largely on Organisation for Economic Co-operation and Development (OECD), African and wider Asian samples, this study is the first to focus specifically on China, where rapid digital growth and health care reforms make the relationship particularly relevant. Third, the study is the first to analyze AI, health expenditure and digital financial inclusion together as determinants of life expectancy within a unified empirical framework using advanced econometric techniques. Finally, this study provides policy implications for China’s health care sector. It highlights the need to strengthen financial inclusion and digital health initiatives to improve access to medical services, enhance population health outcomes, and ensure that technological advancements such as AI are effectively integrated into health care delivery.

Literature Review

Theoretical Review

Zuhair et al. (2024) analyzed the role of AI across various health care specialties. In radiology, AI applications include convolutional neural networks (CNNs), computer-aided detection (CAD), and radiomics for image analysis. In cardiology, AI supports automatic recognition of cardiac structures, prediction of cardiovascular outcomes, and monitoring of interventions and treatment effects. In oncology, it assists in early cancer detection, improves contour precision, classifies tumors, and detects tumor-marker variations. Within intensive care units (ICUs), AI predicts stay durations, readmission probabilities, mortality rates, and identifies patient-ventilator asynchrony. In surgery, AI enhances precision, efficiency, 3D imaging reconstruction, and provides training tools for new surgeons. Finally, in public health, AI accelerates vaccine development and enables rapid data analysis for epidemiological surveillance, especially during epidemics and pandemics. According to Xu et al. (2022), financial structure can influence health outcomes through multiple channels. A lack of financial resources is often cited as a key factor contributing to poor individual health, as high borrowing costs and other constraints prevent people from financing their health needs. Conversely, easy access to financial services allows households and entrepreneurs to expand economic opportunities and manage risks more effectively. The financial structure also impacts economic outcomes without directly affecting consumption of financial services. For example, a more developed financial system can stimulate economic activity, increase labor demand, and generate employment. As incomes rise through economic growth and job creation, individuals can afford better health care services.

From a theoretical perspective, the literature highlights three main approaches to financial inclusion. First, the vulnerable group theory emphasizes that financial inclusion should focus on marginalized segments of society, such as the poor, the young, the elderly, and women. Second, the public goods theory argues that financial inclusion should be available to everyone in society, ensuring that no one is excluded. Finally, the capability theory suggests that financial inclusion expands people’s freedom to make choices regarding essential needs such as quality health care, education, clean water, and sanitation which in turn can improve overall health outcomes (Xiao & Tao, 2022).

The Relationship Between AI and Health

Ronmi et al. (2023) examined the major determinants of life expectancy by analyzing data from 193 countries. Their objective was to distinguish the most influential economic, health, social, immunological, and personal factors affecting longevity. The study used machine learning tree-based models, including extremely randomized trees, applied to a large global data set. Although the exact time period was not specified, the data covered a wide cross-section of countries. The results showed that the extremely randomized tree model produced the most accurate predictions based on mean absolute error (MAE), root mean squared error (RMSE), R², and cross-validation scores. The authors concluded that identifying key determinants can help governments allocate scarce resources more effectively to improve life expectancy. A. F. Khan et al. (2025) analyzed how AI investment affects life expectancy in OECD countries. The study aimed to extend the Grossman health capital model by incorporating AI, governance, and internet usage into the life expectancy framework. Using Generalized Methods of Moment (GMM), the authors examined panel data for 2012–2022. They found that AI investment initially has a negative impact on longevity due to transition costs, but life expectancy increases when AI investment is supported by strong governance and higher internet usage. Gross domestic product (GDP) per capita and out-of-pocket health expenditure showed positive but insignificant effects. The study emphasized the importance of digital inclusion, regulatory frameworks, and institutional quality for achieving long-term gains from AI in health systems. Peddamukkula (2024) explored the role of AI in improving individual life expectancy predictions for the annuity and insurance sector, focusing on a data set of anonymized health and mortality records. The objective was to show how AI can outperform traditional actuarial models. The study used a range of machine learning techniques, including neural networks, decision trees, ensemble models, NLP, and deep learning. The time frame for the underlying health data was not explicitly stated. The findings revealed that AI models significantly improve predictive accuracy compared with actuarial tables, enabling better pricing and risk management in the insurance market. The study highlighted implications for regulation, privacy, and ethical concerns. Yeung and Chung (2025) analyzed AI investment and life expectancy in 25 OECD countries using 226 observations from 2012 to 2024. Drawing on the Grossman model, their objective was to examine how AI adoption interacts with GDP, health care expenditure, governance, and digital access. Using GMM, they reported that past levels of life expectancy have a strong positive effect on current outcomes, reflecting cumulative improvements. Governance quality and internet penetration showed significant positive effects. However, AI investment had a negative and statistically significant impact, suggesting that countries may face transitional disruptions before benefits appear. The authors concluded that strong institutions and digital infrastructure are essential for effective AI integration in health systems. Jadav et al. (2023) focused on improving life expectancy through smart agriculture in the context of food safety. Their objective was to design a blockchain- and AI-based system that detects pesticide overuse in crops. Using performance testing of accuracy, scalability, and latency, the authors evaluated the proposed model against baseline approaches. Although no specific time period was associated with the analysis, the findings showed that the model performs better than alternatives and can help reduce exposure to harmful pesticide levels. The study linked safer agriculture practices to long-term improvements in life expectancy. Ruan et al. (2021) estimated Health-Adjusted Life Expectancy (HALE) for Chongqing, China, using 13.99 million electronic medical records. Their objective was to develop an AI-driven method to generate more accurate and timely HALE estimates. The authors applied NLP, life table techniques, and Sullivan’s method using data from 2017. They reported a life expectancy of 77.9 years and a HALE of 71.7 years, with women experiencing higher HALE than men. Cerebrovascular disease, cancer, and injuries were identified as the largest contributors to years lost due to disability. The study demonstrated the effectiveness of EMR-based AI systems for disease burden assessment and highlighted the potential for broader regional application. Lăzăroiu et al. (2025) examined the evolving research trends in digital technologies that support predictive health care, with a particular focus on the role of digital twins. The authors employed a structured bibliometric analysis combined with qualitative thematic analysis from 2015 to 2025. The findings reveal a growing shift toward integrated and data-driven health care systems, where digital twins serve as a central framework connecting artificial intelligence, machine learning, and internet of Medical Things technologies. The study identifies three emerging thematic areas: integrated patient data ecosystems, predictive and preventive digital twins, and digital twin based treatment planning and patient response simulation. The results indicated an increasing focus on personalized, predictive, and simulation oriented health care models.

The Relationship Between Health Expenditure and Health

Morina et al. (2022) examined how health expenditure (HE) and related socioeconomic factors influence life expectancy in OECD countries. Their objective was to assess the effect of health spending, GDP per capita, productivity, infant mortality, cancer deaths, suicide rates, and potential years of life lost on national longevity. Using a wide set of econometric methods, including linear regression, fixed and random effects, Hausman–Taylor regression, Arellano–Bond GMM, generalized estimating equation (GEE) models, and trend analysis, they analyzed 2005–2018 secondary data from the OECD, International Monetary Fund (IMF), and World Bank. The study found that health expenditure significantly improves life expectancy, confirming a strong and positive relationship between national health care investment and longevity in OECD members. Nkemgha et al. (2021) focused on the impact of public and private HE on LE in Cameroon. The study aimed first to compare the effects of public versus private spending on longevity and second to test causality among these variables. Using ordinary least squares (OLS) regression and the Toda–Yamamoto causality test from 1980 to 2014, the authors found that private HE has a positive and significant effect on LE, while public HE does not. The causality test revealed bidirectional causality between private spending and LE, and a one-way relationship from LE to public spending. Lo et al. (2024) studied how healthy lifestyle factors influence life expectancy and lifetime health care spending in Taiwan. Their objective was to identify the individual and combined effects of five behaviors: nonsmoking, limited alcohol consumption, adequate physical activity, normal weight, and healthy diet. Using cohort data from the National Health Interview Survey and applying a rolling extrapolation algorithm with inverse probability weighting, the study covered a median follow-up of 15.6 years. Results showed that adherence to all five lifestyle factors increased life expectancy by 7.13 years and reduced lifetime health care spending by 28.12%. Nonsmoking, adequate diet, and physical activity had the largest individual effects, highlighting the strong health and cost benefits of healthy behavior. Polcyn et al. (2023) analyzed the effects of HE, environmental pollution, energy use, population size, and income on life expectancy in 46 Asian countries. from 1997 to 2019. The findings showed that higher HE and greater energy consumption improve LE, while CO₂ emissions harm health. Health care spending was the most influential determinant of longevity in Asia, and the results stressed the need for increased health spending and reduced emissions. Giyasova et al. (2025) provided the causality analysis of the connection between health expenditure and LE in Kazakhstan. Using Toda–Yamamoto Granger causality from 2000 to 2021, the study found that HE significantly Granger-causes LE. The authors concluded that sustained increases in health spending are necessary to improve long-term health outcomes in Kazakhstan. Vărzaru (2025) investigated how different types of HE influence health outcomes in the European Union. Using extensive data sets and predictive techniques including artificial neural networks, exponential smoothing, and autoregressive integrated moving average (ARIMA) models, the study analyzed historical trends and future projections. Findings showed that household health care expenditure reduces standardized death rates and improves healthy life years and health expectancy. The study emphasized that strategic, long-term investment—especially household contribution strengthens population health and enhances resilience in EU health care systems.

The Relationship Between Financial Inclusions and Health

Xiao and Tao (2022) examined how financial inclusion (FI) influences population health in Asian countries. Using panel data techniques on data from 2007 to 2019, they analyzed indicators such as internet users, GDP, and foreign direct investment (FDI). The findings showed that both FI and DFI improve LE and reduce death rates. internet use, higher GDP, and foreign investment also contributed positively to population health. The authors concluded that better access to finance strengthens overall health outcomes in Asia. M. D. Alhassan and Adam (2021) studied the global effects of digital inclusion and Information and Communications Technology (ICT) access on the quality of life across 121 countries. Their objective was to understand how ICT access, digital inclusion, and ICT usage collectively shape quality of life. Using partial least squares structural equation modeling (PLS-SEM) and secondary data for 2018, they found that digital inclusion and ICT access significantly improve quality of life, with ICT usage playing a strong mediating role. The study highlighted that improving digital access can enhance well-being worldwide, filling a gap in global-level evidence.

Adediyan (2020) analyzed whether financial inclusion affects life expectancy in 14 West African countries. The study aimed to provide evidence on how access to financial services influences human lifespan. Using dynamic two-step system GMM on panel data from 2010 to 2018, the author controlled for health expenditure, food production quality, electricity access, population, and open defecation. The findings showed strong evidence that financial inclusion raises life expectancy in West Africa, suggesting the need for policies that expand financial access. Naveenan et al. (2024) explored the combined impact of FI, DFI, and health outcomes in developing countries. Their objective was to examine how fintech-based and traditional financial inclusion shape health indicators. Using the Entropy Weight Method to develop a financial inclusion index and applying econometric analysis, the authors analyzed a multi-country data set (time period not explicitly stated but based on recent years). They found that FI improves health outcomes and that digital inclusion strengthens this relationship. The study also showed that ICT adoption can boost both financial inclusion and health performance in emerging markets. Lei et al. (2023) studied the influence of DFI on subjective well-being among the elderly in China. The objective was to assess whether digital finance affects depression, life satisfaction, and self-reported health. Using microdata from the CHARLS survey, covering multiple waves (years implied but not explicitly specified), the authors applied econometric modeling. They found that DFI reduces depression and improves self-reported health, with stronger effects in rural areas and among non-poor groups. The study showed that improvements in digital finance can promote the well-being of vulnerable elderly populations. Malak and Arshad (2024) evaluated how FI affects health care access in 29 developing countries. The objective was to measure how financial services availability contributes to timely and improved health care. Using fixed effects, 2SLS, and system GMM on panel data from 2004 to 2022, they found that financial inclusion significantly improves health care access. Trade openness and FDI also had positive effects, while urbanization had a negative effect on health care access. The authors recommended stronger financial literacy and inclusion programs, especially in rural areas. Wirajing et al. (2024) investigated the effect of financial inclusion on health care in African countries. Covering 2000–2021, their objective was to analyze the role of education as a moderating factor. Using system GMM and multiple health care indicators (life expectancy, immunization rate, and maternal mortality risk), they observed that financial inclusion improves health care outcomes. Education and technology diffusion further enhanced this effect, with education interacting positively with financial inclusion. The results remained robust across different indicators, highlighting the importance of financial literacy and human capital. Immurana et al. (2021) examined how FI affects population health across 33 African countries from 2004 to 2018. Their objective was to provide empirical evidence on the role of financial access in improving life expectancy and reducing death rates. The study used PCA to build a financial inclusion index from ATMs, bank branches, borrowers, and deposits, and applied system GMM for estimation. The findings showed that financial inclusion significantly enhances population health, suggesting that expanding banking services can support better health outcomes across Africa. Based on the literature, the study will test the following hypotheses:

Hypothesis 1 (H1): Artificial intelligence has a positive effect on life expectancy.

Hypothesis 2 (H2): Health expenditure significantly influences life expectancy.

Hypothesis 3 (H3): Digital financial inclusion enhances life expectancy.

This study addresses several important gaps in the existing literature. First, earlier research typically relied on a single measure of AI, such as total AI investment, which limits the ability to fully capture AI development. To address this, the study constructs a more comprehensive AI index using three indicators: AI investment, AI patents, and AI publications, and applies PCA to generate a robust measure of AI progress. Second, while prior studies have examined the influence of AI on health outcomes in OECD countries, Africa, and broader Asian contexts, no study has focused specifically on China, despite its rapid expansion in digital technologies and health care reform. Third, the combined effects of AI, health expenditure, and digital financial inclusion on life expectancy have not been studied together, particularly within China’s unique institutional and technological environment. This study fills these gaps by developing an integrated empirical model that assesses how these three factors jointly influence life expectancy in China, applying advanced econometric techniques to provide reliable evidence.

Methodology and Data

Model Specification

This research aims to analyze the impact of Artificial intelligence, health expenditure and digital Financial inclusion on life expectancy. The model employed in this study is derived from earlier research conducted by literatures:

\begin{array}{l} {LE}_{t} = ϑ_{0} + ϑ_{1} G D P_{t} + ϑ_{2} H E_{t} + ϑ_{3} A I I_{t} + ϑ_{4} D F I_{t} \\ + ϑ_{5} G E_{t} + e_{t} \end{array}

(1)

In Equation 1, LE, GDP, HE, AII, DFI, and GE signifies the gross domestic product, health expenditure, artificial intelligence index, digital financial inclusion index, and government effectiveness, respectively. To avoid data sharpness, reduce the risk of heteroscedasticity, and facilitate easy interpretation of the results, the study transformed the data into natural logarithmic form. Life expectancy and GDP are converted into natural logarithms, while HE, AII, DFI, and GE are kept in their original forms. This is because HE is already expressed in percentage terms, and AII, DFI, and GE contain negative values.

Econometrics Technique

The Augmented Dickey–Fuller (ADF) test is used to determine whether a time series is non-stationary by checking for the presence of a unit root (Dickey & Fuller, 1979). It builds on the standard Dickey–Fuller test by including lagged difference terms to account for autocorrelation. Under this test, the null hypothesis assumes the series contains a unit root, whereas the alternative hypothesis indicates that the series is stationary:

Δ H_{t} = ω + + ϕ t + \partial H_{t - 1} + \sum_{i = 1}^{p} δ_{i} Δ H_{t - i} + u_{t}

(2)

In Equation 2, $Δ H_{t}$ is the first difference of the series, $ω$ is intercept, $H_{t - 1}$ tests for the unit root, and the lagged differences $Δ H_{t - 1}$ . The Phillips–Perron (PP) test also checks for unit roots but uses non-parametric corrections for serial correlation (Phillips & Perron, 1988). It relies on the same base regression as the DF test but adjusts the test statistics instead of adding lagged terms. The null hypothesis assumes non-stationarity, similar to the ADF test.

The autoregressive distributed lag (ARDL) model proposed by Pesaran et al. (2001) is applied to assess the short- and long-run effects of AI, health expenditure, and digital FI on life expectancy in China. This approach is suitable when variables are integrated of order I(0) or I(1). As noted by Duasa (2007), the ARDL framework is appropriate for comparing short- and long-run elasticities in small samples under the usual OLS assumptions, allowing for the detection of cointegration among the variables. Frimpong and Oteng-Abayie (2006) further emphasize that the model performs well when the explanatory variables are integrated of order I(0), I(1), or jointly cointegrated, although it becomes unsuitable when any variable is integrated of order I(2) (Alwago, 2023). To assess both the short-term and long-term equilibrium relationships, Equation 1 was reformulated into the ARDL specification shown in Equation 3:

\begin{array}{l} Δ L E_{t} = ω_{0} + ω_{1} L E_{t - 1} + ω_{2} G D P_{t - 1} + ω_{3} H E_{t - 1} + ω_{4} A I I_{t - 1} \\ + ω_{5} D F I_{t - 1} + ω_{6} G E_{t - 1} + ϕ_{5} e_{t - 1} + \sum_{i = 1}^{n} γ_{1} Δ L E_{t - 1} \\ + \sum_{i = 0}^{n} γ_{2} Δ G D P_{t - 1} + \sum_{i = 0}^{n} γ_{3} Δ H E_{t - 1} \\ + \sum_{i = 0}^{n} γ_{4} Δ A I I_{t - 1} + \sum_{i = 0}^{n} γ_{5} Δ D F I_{t - 1} \\ + \sum_{i = 0}^{n} γ_{6} Δ G E_{t - 1} + e_{t} \end{array}

(3)

The drift component of the model is represented by $ω_{0}$ in Equation 3, the symbol $Δ$ denotes the first difference of the time series, and the error term $e_{t}$ reflects the white noise disturbance. The null hypothesis asserts that there is no long-run relationship among the variables ( $H_{0} : ω_{1} = ω_{2} = ω_{3} = ω_{4} = ω_{5} = = ω_{6} = 0$ ) is evaluated using the F-statistic. When the computed F-value is below the lower bound, Ho is accepted, indicating the absence of cointegration among the variables. After confirming the long-term equilibrium relationship between AI, health expenditure, digital financial inclusion, and life expectancy in China using Equation 7, the study employed the error correction model (ECM) to identify the short-run dynamics. This approach makes it possible to assess the speed at which the system adjusts back to its long-run equilibrium following a short- or medium-term disturbance. The ECM specification presented in Equation 4 is derived directly from the ARDL representation in Equation 3:

\begin{array}{l} Δ L E_{t} = ϕ_{0} + \sum_{i = 1}^{n} γ_{1} Δ L E_{t - 1} + \sum_{i = 0}^{n} γ_{2} Δ G D P_{t - 1} \\ + \sum_{i = 0}^{n} γ_{3} Δ H E_{t - 1} + \sum_{i = 0}^{n} γ_{4} Δ A I I_{t - 1} \\ + \sum_{i = 0}^{n} γ_{5} Δ D F I_{t - 1} + \sum_{i = 0}^{n} γ_{6} Δ G E_{t - 1} \\ + δ E C M_{t - 1} + e_{t} \end{array}

(4)

For the robustness analysis, this study utilized the Quantile regression (QR). The classical regression framework focuses on how an independent variable affects the dependent variable through the conditional mean function $E (y | x)$ . This approach captures only one point of the conditional distribution and offers a limited view of the relationship between variables (Giles, 2018). In practice, distributions of economic variables are often heterogeneous, so the strength and direction of relationships may differ across various points of the outcome distribution. Quantile regression, introduced by Koenker and Bassett (1978), addresses this limitation by estimating effects at different quantiles of the conditional distribution. This provides a more complete understanding of how the explanatory variables influence the dependent variable across its entire distribution (Cade & Noon, 2003), see (Barış-Tüzemen et al., 2020, pp. 20792–20793).

Quantile regression is formally expressed in Equation 5, following the formulation presented by Katchova (2013).

y_{i} = x' β_{q} + e_{i}

(5)

In Equation 5, $β_{q}$ denotes the vector of unknown parameters associated with the $q$ th percentile. While OLS diminishes the sum of squared residuals, $\sum_{i} e_{i}^{2}$ , median regression, also called least absolute deviation (LAD) regression, diminishes the sum of absolute residuals, $\sum_{i} | e_{i} |$ . Quantile regression, in contrast, diminishes a sum that applies asymmetric penalties: $(1 - q) | e_{i} |$ for over-prediction and $q | e_{i} |$ for under-prediction (Baum, 2013). The QR estimator for quantile $q$ is obtained by minimizing the following objective function (Katchova, 2013):

Q (β_{q}) = \sum_{i : y_{i} \geq x_{i}^{'} β} q | y_{i} - x_{i}^{'} β_{q} | + \sum_{i : y_{i} < x_{i}^{'} β} (1 - q) | y_{i} - x_{i}^{'} β_{q} |

(6)

where $0 < q < 1$ . The conditional quantile of $y_{i}$ given $x_{i}$ is expressed as:

Q_{q} (y_{i} x_{i}) = x_{i}^{'} β_{q}

(7)

Quantile regression, which relies on minimizing the absolute values of the residuals, offers more robust estimates than the classical OLS model when the error terms are non-normal or affected by outliers (Giles, 2018), see (Barış-Tüzemen et al., 2020, pp. 20792–20793).

Data

The data has been obtained from the world bank, our world in data, Voronoi and OECD data base. Table 1, shows the Data measurement and sources. Due to the limited number of annual observations from 2013 to 2023, this study converts the data into quarterly frequency (2013Q1 to 2023Q1) to increase the number of observations and improve the efficiency of the estimation. Increasing the sample size helps provide more reliable and robust results in time series analysis, particularly when applying the ARDL model, which performs better with a relatively larger number of observations. This temporal disaggregation technique is commonly used in empirical research when annual data are limited, allowing researchers to capture more dynamic relationships among variables while maintaining the original data structure.

Table 1.

Data Measurement and Sources.

Symbols	Variables	Measurement	Data sources
LE	Life expectancy	Life expectancy at birth, total (years)	World bank
GDP	Gross domestic product	GDP per capita (constant 2015 US$)	World bank
HE	Health expenditure	% of GDP	World bank
AII	Artificial intelligence index	Index	Our world in Data, OECD, Voronoi
DFI	Digital financial inclusion	Index	World bank
GE	Government effectiveness	Estimate	World bank

https://www.voronoiapp.com/technology/China-Has-Accumulated-70-of-the-Worlds-AI-Patents-4740

Table 2, shows the PCA output. The PCA results show that the first component (Comp1) has an eigenvalue of 2.176, explaining 72.5% of the total variance, while the second and third components add little additional explanatory power (23.3% and 4.1%, respectively). This means that the first component captures most of the information from the original AI variables. From the eigenvectors, all three indicators have strong positive loadings on Comp1:AI patents (AIP): 0.653 Private AI investment (AIIV): 0.530 and AI research publications (AIRP): 0.54 Since the loadings are positive and similar in magnitude, the first principal component can be interpreted as a comprehensive AI development index, representing overall progress in AI patents, investments, and research. For digital financial inclusion, the first component (Comp1) has an eigenvalue of 2.35, explaining 78.3% of the total variance. The remaining components contribute very little, so only Comp1 is retained for constructing the index. The eigenvector loadings for Comp1 are: Automated teller machines per 100,000 adults (ATM): 0.585, Commercial bank branches per 100,000 adults (CBB): 0.640, Depositors with commercial banks (% of adults; DCC): 0.498 and Financial inclusion is a composite index of ATMs (per 100,000 adults), bank branches (per 100,000 adults), and outstanding deposits with commercial banks (% of GDP). Following the PCA approach, the AI and DFI indices were developed from the first principal component (PC1), which explains the largest share of the total variance in the selected indicators.

Table 2.

PCA Output for AI and Digital Financial Inclusions Index.

Artificial intelligence index
Component	Eigenvalue	Difference	Proportion	Cumulative
Comp1	2.176	1.476	0.725	0.725
Comp2	0.7	0.576	0.233	0.959
Comp3	0.124		0.041	1
Principal components (eigenvectors)
Variable	Comp1	Comp2	Comp3
AIP	0.653	−0.013	−0.757
AIIV	0.53	0.722	0.445
AIRP	0.541	−0.692	0.478
Digital financial inclusion index
Component	Eigenvalue	Difference	Proportion	Cumulative
Comp1	2.35	1.743	0.783	0.783
Comp2	0.607	0.564	0.202	0.986
Comp3	0.043		0.014	1
Principal components (eigenvectors)
Variable	Comp1	Comp2	Comp3
ATM	0.585	−0.545	0.601
CBB	0.64	−0.144	−0.755
DCC	0.498	0.826	0.265

Results and Discussions

Results

Table 3 presents the summary statistics for the main variables used in the study, including LE, GDP, HE, AII, DFI, and GE. The mean and median values of LE (4.35 and 4.35) are very close, indicating a symmetrical distribution. GDP and HE also show minimal difference between their mean and median values, suggesting relatively stable growth over the sample period. The AII and DFI indices have means close to zero due to normalization during the PCA process, while their wide ranges (AII: –1.61 to 1.99; DFI: –3.28 to 1.22) reflect considerable variation across time or countries in AI adoption and financial inclusion levels. The Jarque–Bera statistics for all variables are low, and their associated probabilities exceed 0.10, confirming that all series are approximately normally distributed.

Table 3.

Descriptive Statistics.

Variables	Mean	Median	Maximum	Minimum	Jarque-Bera	Probability
LE	4.3503	4.3530	4.3593	4.3367	1.0510	0.5913
GDP	9.1723	9.1900	9.4322	8.8816	0.6823	0.7110
HE	5.1856	5.1679	5.7000	4.7149	0.5889	0.7449
AII	0.0000	0.0366	1.9933	−1.6131	1.1805	0.5542
DFI	0.0000	0.8610	1.2194	−3.2808	3.0878	0.2135
GE	0.4734	0.4942	0.8079	−0.0051	1.0254	0.5989

Table 4, reports the results of the ADF and PP tests for stationarity. Both tests were applied at levels I(0) and first differences I(1) to determine the integration order of each variable. The results show that LE, GDP, HE, and AII are non-stationary at levels but become stationary after first differencing, indicating they are integrated of order one, I(1). In contrast, DFI and GE are stationary at level in at least one of the tests, suggesting they are I(0). The consistency between the ADF and PP results confirms the robustness of the findings. The mix of I(0) and I(1) variables justifies the application of the ARDL bounds testing approach for cointegration analysis in the next stage of the study.

Table 4.

Unit Root Test.

	ADF		PP
Variables	I(0)	I(1)	I(0)	I(1)
LE	−2.218	−5.534*	−2.377	−5.383*
GDP	−2.186	−4.747*	−2.285	−4.747*
HE	−0.566	−4.081*	0.135	−3.229*
AII	−0.013	−4.403*	−0.032	−4.769*
DFI	−3.963*	−3.419*	−6.233*	−3.068*
GE	−2.920***	−4.562*	−2.920***	−8.313*

Note, *,** & *** represent the 1%, 5% & 10% significance level.

Table 5 presents the results of the ARDL bounds test for the existence of a long-run relationship among the variables. The calculated F-statistic value is 10.7529, which is greater than the upper critical bound (I(1)) at all conventional significance levels (1%, 2.5%, 5%, and 10%). This confirms the rejection of the null hypothesis of no cointegration, indicating the presence of a long-run equilibrium relationship among LE, GDP, HE, AII, DFI and GE.

Table 5.

ARDL F-Bounds Test.

Test Statistic	Value	Significance (%)	I(0)	I(1)
F-statistic	10.7529	10	1.81	2.93
k	5	5	2.14	3.34
		2.50	2.44	3.71
		1	2.82	4.21

Table 6 reports the estimated long-run and short-run coefficients of the ARDL model, examining the effects of health expenditure, artificial intelligence, digital financial inclusion on LE. In the long run, all variables are positively and significantly associated with life expectancy. Specifically, GDP (0.459) and health expenditure (0.383) have the strongest effects, indicating that higher economic growth and greater investment in health care contribute substantially to improved longevity. The coefficients of AII (0.060), DFI (0.145), and GE (0.108) are also positive and statistically significant, suggesting that advancements in artificial intelligence, expansion of digital financial inclusion, and better governance all play important roles in enhancing health outcomes. In the short run, GDP (0.076), HE (0.507), AII (0.029), and GE (0.051) remain positive and significant, showing that improvements in these factors have immediate beneficial effects on life expectancy. The coefficient of DFI (0.093) is positive and weakly significant (p = 0.084), implying a moderate short-term effect. The error correction term (ECM₍₋₁₎) is negative and statistically significant (–0.278, p = 0.006), confirming that any short-run disequilibrium adjusts toward long-run equilibrium at a speed of about 27.8% per year, indicating the model is stable and converges to its long-run path after short-term shocks.

Table 6.

ARDL Estimates.

Variable	Long run			Short run
Variable	Coefficient	SE	Prob.	Coefficient	SE	Prob.
GDP	0.459*	0.024	.000	0.076*	0.009	0.000
HE	0.383*	0.052	.000	0.507*	0.017	0.000
AII	0.060*	0.002	.000	0.029*	0.007	0.000
DFI	0.145*	0.011	.000	0.093***	0.048	0.084
GE	0.108*	0.005	.000	0.051*	0.002	0.000
C	4.751*	0.209	.000	—	—	—
ECM(–1)	—	—	—	–0.278	0.095	0.006

Note, *,** & *** represent the 1%, 5% & 10% significance level.

Table 7 summarizes the post-estimation diagnostic tests conducted to ensure the reliability of the ARDL model. The Jarque–Bera test statistic (1.057, p = 0.510) confirms that the residuals are normally distributed (see Figure 2). The Breusch–Godfrey LM test (F = 0.054, p = 0.946) shows no serial correlation, indicating that the model’s residuals are independent over time. Similarly, the Autoregressive Conditional Heteroskedasticity (ARCH) test (F = 0.743, p = 0.393) suggests the absence of heteroskedasticity, meaning the variance of the residuals is constant. Finally, the Ramsey RESET test (F = 0.802, p = 0.739) verifies that the model is correctly specified with no evidence of omitted variable bias or misspecification. In Figures 3 and 4, the Cumulative Sum (CUSUM) and CUSUM of Squares (CUSUMSQ) tests indicate that all coefficients are stable over time, as the plotted blue line remains within the 5% critical boundaries (represented by the red lines). This confirms the structural stability of the ARDL model throughout the sample period. Finally, the diagnostic and stability results demonstrate that the estimated model is econometrically sound, reliable, and suitable for inference and policy interpretation.

Table 7.

ARDL Diagnostic Analysis.

Tests	F statistics	P-value	Remarks
Jarque-Bera	1.057	0.510	Normal
Breusch-Godfrey LM	0.054	0.946	No serial correlation
ARCH	0.743	0.393	No heteroskedasticity
Ramsey RESET	0.802	0.739	Model is correctly specified

Figure 2.

Normality test.

Figure 3.

CUSUM test.

Figure 4.

CUSUM square test.

Quantile regression and fully modified ordinary least squares (FMOLS) were used to test the robustness of the ARDL results across the distribution of life expectancy. In Table 8, shows the estimates of quantile regression, HE, AII, and DFI show consistently positive and significant effects across all quantiles, confirming their strong role in improving life expectancy. GDP has a mixed effect, becoming more positive and significant at higher quantiles, while GE is mostly positive, with minor variation at the middle quantiles. These results indicate that the ARDL findings are robust, with key variables maintaining their influence across different levels of life expectancy. According to the FMOLS results, GDP, HE, AII, DFI, and GE have a positive effect on life expectancy, which is consistent with the findings of the ARDL model.

Table 8.

Robustness Analysis.

Quantile regression
	Fifth	10th	20th	30th	40th	50th	60th	70th	80th	90th
GDP	−0.034	−0.034	−0.034*	0.034*	0.044**	0.043*	0.018*	0.020	0.035*	0.033**
HE	0.005*	0.005*	0.005*	0.005**	0.008*	0.005*	0.006*	0.003*	0.004**	0.004***
AII	0.005*	0.005*	0.005*	0.005*	0.005*	0.006*	0.003*	0.001*	0.001*	0.001**
DFI	0.004*	0.004*	0.004*	0.004*	0.004*	0.005*	0.002*	0.001**	0.001**	0.001*
GE	0.005**	0.005**	0.005*	0.005*	−0.004*	0.003*	0.001*	0.003*	0.002*	0.001**
C	4.634*	4.634*	4.634*	4.634*	4.713*	4.725*	4.161	4.149*	4.009*	4.033*
FMOLS
Variable	Coefficient	SE	Prob.
GDP	0.546*	0.028	0.000
HE	0.126**	0.050	0.039
AII	0.033*	0.007	0.003
DFI	0.108*	0.029	0.009
GE	0.749*	0.120	0.000

Note. *, **, and *** represent the 1%, 5%, and 10% significance level. Dependent variable: life expectancy.

Discussion

The finding shows that GDP has positive effect on LE, higher GDP increases national income, enabling greater public and private spending on health services. It improves nutrition, sanitation, and access to clean water, while supporting better infrastructure, hospitals, and health care technology. Moreover, Economic growth allows investment in preventive care and education, reducing mortality rates and enhancing overall health outcomes. The finding is consistent with the finding of Azam et al. (2023), and G. N. Alhassan et al. (2025), who found that Per capita income has positive effect on LE in Pakistan, while GDP per capita enhances LE across all income levels in developing economies.

Health expenditure has positive effect on LE, increase in health expenditure provides more resources for hospitals, clinics, and medical staff. It allows for better medicines, equipment, and vaccination programs, while higher HE improves preventive and curative care, reducing disease burden. It ensures that both rural and urban populations have access to quality health care. The results are same with the finding of Uddin et al. (2024), who examined the determinants of LE, they found that HE has positive effect on LE. Nkemgha et al. (2021) found that private HE has a positive and significant effect on LE, while public HE does not. The causality test revealed bidirectional causality between private spending and LE, and a one-way relationship from LE to public spending. Zhang et al. (2026) examined the impact of ICT, health expenditure, and institutional quality on LE in China. The finding showed that ICT, health expenditure and institutional quality have positive effect on LE in China.

AI has positive effect on LE, the reason of the positive sign in that, AI improves diagnostics through predictive algorithms and early disease detection, helping to identify and treat illnesses sooner, thereby reducing mortality and enhancing LE. It enhances hospital management, patient monitoring, and personalized treatments, while supporting health care research and drug development efficiency, which leads to better health outcomes. It reduces medical errors and optimizes resource allocation in health care systems, ensuring more effective care and ultimately contributing to longer LE. A. F. Khan et al. (2025) reported that AI investment initially has a negative impact on longevity due to transition costs, but LE increases when AI investment is supported by strong governance and higher internet usage. Yeung and Chung (2025) found that AI investment had a negative and statistically significant impact, suggesting that countries may face transitional disruptions before benefits appear.

Digital financial inclusion has positive effect on LE, DFI enables easier access to health insurance and financial services for all populations, allowing timely treatment and reducing mortality, thereby improving LE. It allows timely payment for health care, reducing treatment delays, while digital financial tools facilitate health care subsidies and emergency funds, supporting faster and more effective care. Inclusion promotes equitable access, particularly for vulnerable or rural groups, ensuring broader health coverage and contributing to higher LE. Xiao and Tao (2022) showed that both FI and DFI improve LE and reduce death rates. Naveenan et al. (2024) explored the combined impact of FI, digital inclusion on health outcomes in developing countries. They found that FI improves health outcomes and that digital inclusion strengthens this relationship. The study also showed that ICT adoption can boost both financial inclusion and health performance in emerging markets.

Government effectiveness has positive effect on LE, effective governance ensures proper allocation and oversight of health care resources. It strengthens policy implementation, monitoring, and health program delivery, while Government effectiveness reduces corruption, ensuring funds reach intended health projects. It promotes accountability, transparency, and responsiveness in the health sector. The finding is consistent with the finding of K. Khan et al. (2024), who analyzed the government effectiveness and LE nexus, they found that government effectiveness has positive effect on LE in Pakistan. Bunyaminu et al. (2022) found that health expenditure has a positive and significant impact on LE. However, when combined with the moderating effect of government effectiveness, health expenditure appears to reduce LE.

Conclusion and Policy Recommendation

This study examined the impact of Artificial intelligence, health spending, and digital financial inclusion on LE in China from 2013Q1 to 2023Q4. The analysis employed unit root test, ARDL and quantile regression. The findings indicated that GDP, AI, HE, DFI, and government effectiveness all have positive effects on LE. The results highlight that AI enhances diagnostics, treatment efficiency, and health care resource allocation, contributing to longer LE. Health expenditure strengthens health care infrastructure, preventive care, and medical service delivery. DFI ensures equitable access to health care. Government effectiveness supports proper allocation and oversight of health resources.

Based on the findings, this study proposes several policy recommendations for the Chinese government to improve life expectancy. The government should invest in AI-powered diagnostic technologies that enable the early detection of diseases. Early diagnosis can reduce mortality rates and improve treatment outcomes, which ultimately contributes to higher life expectancy. Hospitals should adopt AI-based patient monitoring systems and personalized treatment approaches to ensure timely medical interventions and more efficient health care delivery. Greater support should be provided for AI-driven health care research and drug development, as faster medical innovation can significantly improve public health and longevity.

In addition, increasing public spending on health care infrastructure is essential. Expanding and upgrading hospitals and clinics can improve access to medical services and help reduce the overall burden of disease. The Chinese government should also allocate more resources to preventive health care programs, including vaccination campaigns and public health education, as these initiatives can lower the incidence of chronic and infectious diseases and promote longer lives. Furthermore, investing in the training and recruitment of qualified medical professionals is important to ensure the availability of high quality health care services.

Digital financial inclusion can also play an important role in improving health outcomes. Expanding access to digital payment systems and mobile banking for health care services can make it easier for individuals to pay for medical treatment on time, reducing delays in care. Similarly, promoting digital health insurance and savings platforms can help households better manage medical expenses and access essential health care services when needed. Special attention should also be given to vulnerable and rural populations through digital financial literacy programs, enabling them to effectively use financial services and benefit from improved health care access. Finally, strengthening governance and accountability in the allocation of health care resources is crucial. Transparent and efficient distribution of funds to hospitals and clinics can enhance the quality of health care services. At the same time, the government should ensure the effective implementation and monitoring of evidence-based health policies. Strong policy enforcement can improve preventive care, reduce mortality, and ultimately support higher life expectancy.

This study has several limitations that also offer directions for future research. First, the study considers only a limited number of factors, including GDP, health expenditure, artificial intelligence, digital financial inclusion, and government effectiveness, while other important economic, demographic, and social determinants of life expectancy are not included. Future studies should incorporate additional variables to better capture the multidimensional nature of health outcomes. Second, due to the lack of precise data on AI-related health patents, this study was unable to include a fully accurate measure of AI adoption in the health care sector. Future research may address this limitation by incorporating more comprehensive indicators of AI in health care once such data become available. Third, this study developed a digital financial inclusion index using three indicators due to the unavailability of other financial data. Future studies should incorporate additional indicators if they become available. Fourth, the analysis focuses solely on China, which may limit the broader applicability of the findings. Future studies could extend the analysis to other countries or conduct cross-country comparisons to examine whether similar relationships exist in different economic and institutional settings.

Footnotes

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

AI Declaration

The author(s) declare that Gen AI was used in the creation of this manuscript. Artificial intelligence (AI) tools, including ChatGPT, and language enhancement software such as Grammarly were used in the preparation of this manuscript for tasks such as editing and language improvement. All AI-assisted content was carefully reviewed and revised to ensure accuracy, originality, and compliance with ethical standards. The final responsibility for the manuscript remains with the authors.

References

Acheampong

A. O.

Tetteh

G. K.

(2024). Does financial inclusion matter to population health? Insight from a global dataset. Social Indicators Research, 172(3), 1005–1040.

Adediyan

A. R.

(2020). Is financial inclusion a determinant of life expectancy? Evidence from West African Region. Studies in Business and Economics, 23(1), 22–37.

Alhassan

G. N.

Bekun

F. V.

Rashidi

Karalinç

(2025). Multidimensional determinants of life expectancy in developing countries: Evidence from Preston curve hypothesis. Review of Development Economics, 29(4), 2736–2746.

Alhassan

M. D.

Adam

I. O.

(2021). The effects of digital inclusion and ICT access on the quality of life: A global perspective. Technology in Society, 64, 101511.

Alwago

W. O.

(2023). The nexus between health expenditure, life expectancy, and economic growth: ARDL model analysis for Kenya. Regional Science Policy & Practice, 15(5), 1064–1086.

Azam

Uddin

Saqib

(2023). The determinants of life expectancy and environmental degradation in Pakistan: Evidence from ARDL bounds test approach. Environmental Science and Pollution Research, 30(1), 2233–2246.

Barış-Tüzemen

Ö.

Tüzemen

Çelik

A. K

. (2020). Does an N-shaped association exist between pollution and ICT in Turkey? ARDL and quantile regression approaches. Environmental Science and Pollution Research, 27(17), 20786–20799.

Baum

C. F.

(2013). Quantile regression. http://fmwww.bc.edu/EC-CS. http://fmwww.bc.edu/EC-C/S2013/823/EC823.S2013.nn04.slides.pdf

Beylik

Cirakli

Cetin

Ecevit

Senol

(2022). The relationship between health expenditure indicators and economic growth in OECD countries: A Driscoll-Kraay approach. Frontiers in Public Health, 10, 1050550.

10.

Bunyaminu

Mohammed

Yakubu

I. N.

Shani

Abukari

A. L.

(2022). The effect of health expenditure on average life expectancy: Does government effectiveness play a moderating role? International Journal of Health Governance, 27(4), 365–377.

11.

Cade

B. S.

Noon

B. R.

(2003). A gentle introduction to quantile regression for ecologists. Frontiers in Ecology and the Environment, 1(8), 412–420.

12.

Daxue Consulting. (2025, January 13). AI in Chinese healthcare: From medical imaging to AI hospitals. https://daxueconsulting.com/ai-healthcare-china/

13.

Dickey

D. A.

Fuller

W. A.

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366a), 427–431.

14.

Duasa

(2007). Determinants of Malaysian trade balance: An ARDL bound testing approach. Global economic review, 36(1), 89–102.

15.

Eurostat. (2025). Life expectancy. Eurostat, Statistics Explained. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Glossary:Life_expectancy#:~:text=Enter%20search%20term-,Life%20expectancy,females%20and%20the%20total%20population

16.

Frimpong

J. M.

Oteng-Abayie

E. F.

(2006). Bivariate causality analysis between FDI inflows and economic growth in Ghana. https://mpra.ub.uni-muenchen.de/351/

17.

Giles

(2018). Review of handbook of quantile regression. https://www.dropbox.com/scl/fi/77x5m6y18h64uh1s9lj52/Handbook-of-Quantile-Regression.pdf?rlkey=mdyotux4qydh0kzrznkf0oxp2&dl=0

18.

Giyasova

Eyvazova

Mahmudova

Guliyev

Vasa

(2025). Examining the impact of health expenditure on life expectancy in Kazakhstan: A causality analysis. International Journal of Safety and Security Engineering, 15(6), 1159–1164.

19.

Immurana

Iddrisu

A. A.

Boachie

M. K.

Dalaba

M. A.

(2021). Financial inclusion and population health in Africa. Journal of Sustainable Finance & Investment. Advance online publication. https://doi.org/10.1080/20430795.2021.1953929

20.

Jadav

N. K.

Rathod

Gupta

Tanwar

Kumar

Alkhayyat

(2023). Blockchain and artificial intelligence-empowered smart agriculture framework for maximizing human life expectancy. Computers and Electrical Engineering, 105, 108486.

21.

Katchova

(2013). Quantile regression. https://docs.google.com/file/d/0BwogTI8d6EEiYmVzN0kyQ19CU3M/edit?resourcekey=0-j7BdwyCNljyvfRBVl5mR4g

22.

Kelly

C. J.

Karthikesalingam

Suleyman

Corrado

King

(2019). Key challenges for delivering clinical impact with artificial intelligence. BMC Medicine, 17(1), 195.

23.

Khan

A. F.

Hassan

Din

S. U.

(2025). Integrating artificial intelligence into life expectancy model: An analysis of OECD countries. Journal of Asian Development Studies, 14(2), 11–26.

24.

Khan

Zeeshan

Moiz

Bano

Khan

M. H.

Ahmad

Ateya

A. A.

(2024). Influence of government effectiveness, health expenditure, and sustainable development goals on life expectancy: Evidence from time series data. Sustainability, 16(14), 6128.

25.

Koenker

Bassett

Jr. (1978). Regression quantiles. Econometrica: Journal of the Econometric Society, 46(1), 33–50.

26.

Lăzăroiu

Szpilko

Gedeon

Halicka

Rzepka

(2025). Digital technologies supporting predictive healthcare: Review of research trends. Human Technology, 21(3), 694–730. https://doi.org/10.14254/1795-6889.2025.21-3.10

27.

Lei

Shen

Yang

(2023). Digital financial inclusion and subjective well-being—Evidence from China health and retirement longitudinal study. China Economic Review, 81, 102013.

28.

W. C.

T. H.

Shih

C. Y.

Lin

H. H.

Hwang

J. S.

(2024). Impact of healthy lifestyle factors on life expectancy and lifetime health care expenditure: Nationwide cohort study. JMIR Public Health and Surveillance, 10(1), e57045.

29.

Malak

Arshad

(2024). Impact of financial inclusion on healthcare access: Evidence from developing countries. Journal of Economic and Administrative Sciences. Advance online publication. https://doi.org/10.1108/JEAS-02-2024-0057

30.

Mohammed

M. A.

Mohammed

V. A.

Mohammed

V. A.

(2022). Impact of artificial intelligence on the automation of digital health system. International Journal of Software Engineering & Applications, 13(6), 23–29.

31.

Morina

Komoni

Kilaj

Selmonaj

Grima

(2022). The effect of health expenditure on life expectancy. International Journal of Sustainable Development and Planning, 17(5), 1389–1401.

32.

Naveenan

R. V.

Liew

C. Y.

Kijkasiwat

(2024). Nexus between financial inclusion, digital inclusion and health outcomes: Evidence from developing economies. Social Indicators Research, 174(1), 367–408.

33.

Nkemgha

G. Z.

Tékam

H. O.

Belek

(2021). Healthcare expenditure and life expectancy in Cameroon. Journal of Public Health, 29(3), 683–691.

34.

Oniore

J. O.

Aiyedogbon

J. O.

Aigba

(2024). Public health spending indicators and life expectancy in Nigeria: 1990-2023. Medical and Health Sciences European Journal, 8(6), 53–75.

35.

Ortiz-Ospina

(2017). Life expectancy: What does this actually mean? Our World in Data. https://ourworldindata.org/life-expectancy-how-is-it-calculated-and-how-should-it-be-interpreted

36.

Peddamukkula

P. K.

(2024). Artificial Intelligence in Life Expectancy Prediction: A Paradigm Shift for Annuity Pricing and Risk Management. International Journal of Computer Technology and Electronics Communication, 7(5), 9447–9459.

37.

Pesaran

M. H.

Shin

Smith

R. J.

(2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), 289–326.

38.

Phillips

P. C.

Perron

(1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346.

39.

Polcyn

Voumik

L. C.

Ridwan

Ray

Vovk

(2023). Evaluating the influences of health expenditure, energy consumption, and environmental pollution on life expectancy in Asia. International Journal of Environmental Research and Public Health, 20(5), 4000.

40.

Rajpurkar

Chen

Banerjee

Topol

E. J.

(2022). AI in health and medicine. Nature Medicine, 28(1), 31–38.

41.

Ronmi

A. E.

Prasad

Raphael

B. A.

(2023). How can artificial intelligence and data science algorithms predict life expectancy—An empirical investigation spanning 193 countries. International Journal of Information Management Data Insights, 3(1), 100168.

42.

Ruan

Jin

Deng

Liu

Xie

Long

Yuan

Guo

Huang

Yang

Han

Zhuang

. . .Xu

(2021). Health-adjusted life expectancy (HALE) in Chongqing, China, 2017: An artificial intelligence and big data method estimating the burden of disease at city level. The Lancet Regional Health–Western Pacific, 9, 100110.

43.

Tay

L. Y.

Tai

H. T.

Tan

G. S.

(2022). Digital financial inclusion: A gateway to sustainable development. Heliyon, 8(6), e09766.

44.

Uddin

Khan

M. A.

Tariq

Khan

Malik

Z. K.

(2024). Revisiting the determinants of life expectancy in Asia—Exploring the role of institutional quality, financial development, and environmental degradation. Environment, Development and Sustainability, 26(5), 11289–11309.

45.

Vărzaru

A. A.

(2025). Assessing the Relationships of Expenditure and Health Outcomes in Healthcare Systems: A System Design Approach. Healthcare (Basel). 13(4), 352.

46.

Wirajing

M. A. K.

Haruna

Nchofoung

T. N.

(2024). Financial inclusion and healthcare in Africa: Examining the moderating role of education. Review of Development Economics, 28(1), 97–127.

47.

Xiao

Tao

(2022). Financial inclusion and its impact on health: Empirical evidence from Asia. Frontiers in Public Health, 10, 948964.

48.

Tao

Mahmood

C. K.

Altuntaş

(2022). Re-examining the financial structure and health nexus in Asian economies. Frontiers in Public Health, 10, 860325.

49.

Yeung

Chung

(2025). Technology, Institutions, and Longevity: An Empirical Analysis of AI Investment and Life Expectancy in the OECD. Journal of Policy Options, 8(3), 17–29.

50.

Zhang

Shao

Uddin

Zhang

(2026). Do environmental degradation, ICT, health expenditure, and institutional quality affect life expectancy? American Journal of Men’s Health, 20(1), 15579883261420408.

51.

Zuhair

Babar

Ali

Oduoye

M. O.

Noor

Chris

Rehman

L. U.

(2024). Exploring the impact of artificial intelligence on global health and enhancing healthcare in developing nations. Journal of Primary Care & Community Health, 15, 21501319241245847.