Introduction

The Characteristics of Economic Progress have changed dramatically throughout time. Industrialization, modernization, and economic expansion drove development after WWII, with technical advances contributing to greater worker productivity and structural transformation until the 1970s (Skare and Riberio Soriano, 2021). Emerging economies liberalized in order to catch up with established economies. Policymakers supported low inflation and unpredictability in growth from the 1980s through the 2000s. Still, the dot-com crash, Industry 4.0, and the 2008 financial crisis spurred a rethinking of development strategies toward goal-based methods. Digital technology has been a driving factor in the fourth industrial revolution, influencing economic growth in the twenty-first century (Ha, 2022a; Ha et al., 2023; Ha and Thanh, 2022).

A fundamental change is taking place in Vietnam as a result of digital transformation. Nevertheless, Vietnam is still struggling to achieve its digitalization goals. However, the Vietnamese government is developing plans to overcome these challenges in order to transform the economy of Vietnam completely. Furthermore, the Ministry of Information and Communication in Vietnam has proposed that the digital technology industry lead the implementation of Industry 4.0 in the country, with revenues growing 2.5 times faster than the overall GDP. In 2023 and 2025, the digital technology sector is expected to contribute between 6 and 6.5% of the overall GDP growth, with ICT exports amounting to 137,000 million dollars and 160,000 million dollars, respectively. The ICT sector in Vietnam is estimated to generate $148 billion in revenue in 2022, an increase of 8.7% over last year. The industry contributed $34.3 billion to GDP in 2022, an increase of 8.7%. Exports of hardware and electronic components increased by 11.6% to $136 billion in 2022, resulting in a trade surplus of $26 billion, up significantly from $4 billion in 2021. There are 70,000 digital tech firms in the United States as of 2022, an increase of 9.6% over the previous year, according to the MIC report. It is estimated that approximately 60% of these firms are shifting from low-value to high-value production processes. Furthermore, Vietnam is expected to become a leader in digital technology and develop new products and services related to artificial intelligence, the Internet of Things, big data, 5G or even 6G networks, and next-generation networks.

It is evident that the Vietnamese government has placed a high priority on its digitalization target along with a major transformation of its economic system. Human resources, however, are critical to achieving this goal. Vietnam faces the challenge of raising labor productivity to a new level in pursuing this goal. The MIC report indicates that digital skills should be developed at a higher level in the workforce. By 2023, it is estimated that there will be a shortage of up to one million information and communication technology engineers in Vietnam due to a lack of adequate information and communication technology capabilities. According to the World Bank, only 40% of Vietnamese businesses have sufficient information and communication technology capabilities. In order to transform Vietnam’s economy into a digital economy, it may be necessary to develop human capital to increase Vietnam’s competitiveness in the global market.

In the literature, some scholars investigate the one-sided relationship between globalization and digitalization. For example, Miranda et al. (2016) discovered a nonlinear method for embracing or rejecting a technical solution or innovation. Straub (2009) investigated adoption and dissemination and discovered numerous elements that impact technology uptake. Because of the current complexity of information technology, adoption models must adapt quickly. According to Straub (2009), the adoption of technology is a complex social and developmental process influenced by human attitudes. Globalization facilitates the spread of foreign knowledge and increases worldwide rivalry. Fatima (2017) investigates the relationship between adopting new technologies and openness at the company level using data from 30 industrialized and emerging nations. According to the study, foreign direct investment (FDI) is less likely to innovate in process innovation than local firms. Despite the potential benefits of globalization in terms of information dissemination, not all nations and organizations are able to capitalize on these opportunities. As Archibugi and Pietrobelli (2003) point out, developing countries are not immune to this problem. Our research aims to look at the influence of globalization on adopting new digital technologies and how it relates to other technology adoption processes. Although earlier research has looked at globalization’s impact on technology uptake on a national and regional basis, this study looks at the link, with a focus on the case of developing countries. We employed a typical model to examine how globalization’s economic, social, and political components impact technological adoption.

Hence, this study looks at how globalization influences the rate at which digital technology is embraced globally and vice versa. The term “globalization” relates to the increasing interconnectivity of markets and multinational businesses’ cross-border activity. Globalization as a process that encompasses the causes, courses, and outcomes of transnational and transcultural integration of human and nonhuman activities (Al-Rodhan, 2006). Digitalization, on the other hand, refers to the incorporation of digital technology into many elements of daily life (Skare and Riberio Soriano, 2021). Because globalization and digitalization have increased in recent years, it is acceptable to investigate if there is a causal relationship between them. We examine the key components that drive the transfer of digital technologies to highlight the significance of globalization in this process. We collect data at the national level on many indices, including the digitalization captured by Internet users and mobile subscriptions, an openness to the global economic and financial network, to isolate the influence of economic globalization. We employ time-series data from a sample of Vietnam to examine this link. Our findings suggest that globalization promotes creativity and speeds technological transfer.

As a result of globalization, Vietnam is experiencing more dynamic times than ever before (Chi and Dan, 2024; Ha, 2023; Hung et al., 2022). Due to the global restructuring of manufacturing industries, Vietnam has experienced unprecedented economic growth over the last three decades (Abbas et al., 2024; Balsa-Barreiro et al., 2019b; Ha, 2023b; Thai Hung, 2022). Driven by the globalization process, it has been common for transnational companies headquartered in developed countries to move their manufacturing sections to developing countries where they can benefit from lower labor costs, cheaper land prices, and almost non-existent environmental awareness (Balsa-Barreiro et al., 2019b). This study examined transportation systems around the world, implying a different understanding of the geographic space (Balsa-Barreiro et al., 2019a; Balsa-Barreiro et al., 2019b). As a result of this high degree of connectivity, goods and people are more closely located, resulting in increased opportunities and wealth as potential markets are expanded (Pentland, 2014). Economic globalization has resulted in a redistribution of wealth across the globe, with developing countries accounting for a greater share of the global economy than developed countries. As a result of economic globalization, energy demands, population, and particularly the urban population are also redistributed globally (Balsa-Barreiro et al., 2019b). Vietnam’s economic growth is compensated by the high demands on energy and the sustainability of the environment (Do and Nguyen, 2020; Ha, 2023b, 2023a). As a result of globalization, Figge et al. (2017) have studied the impact of ecological footprints on the environment. There has been a significant increase in energy demand in Vietnam, as evidenced by the fact that more rapid economic growth is typically associated with inefficient energy use (Hien, 2019; Nguyen et al., 2024). There is evidence that Vietnam could have generated carbon dioxide at a faster rate over the past few decades than developed countries (Ward and Mahowald, 2014).

Given that the literature suggests that digitalization has an impact on a variety of economic factors, we anticipate that there will be dynamic relationships between economic globalization and digitalization in Vietnam. Many studies, including Devold et al. (2017), Herzog et al. (2017), and Pop (2020), have suggested that the industrialization process can advance more quickly when using digital information systems to provide lower product intermediate and labor costs. Additionally, the global business practices of companies have been impacted by digitalization (Dethine et al., 2020). By lowering barriers to international trade, digitization creates new investment opportunities (Damgaard et al., 2018). The release of digital fundraising platforms, which allow investors to receive direct and indirect benefits without using conventional intermediaries, can revolutionize foreign investments through digitalization (Damgaard et al., 2018). The growth of digital information systems may increase access to new markets. Economic agents can purchase financial services at a cheaper cost and with more efficiency as a result of decreased information flow costs, which lowers the overall expenses of economic activity. Digitalization can thereby boost the competitiveness of the goods and services provided by multinational corporations, raising both the GDP of the host country and the country where the multinational corporation is based. Ha (2022b) employs a combination of econometric strategies and empirical methods using a global sample of 27 European nations from 2015 to 2019 to demonstrate that digitalization is positively correlated with FD. Digitalization has increased the efficiency and depth of financial markets and institutions, yet access to both has diminished.

The main goal of this study is to examine how the openness level to the global economic and financial network and digitalization in Vietnam changed from 1996 to 2019. As far as we are aware, this is the very first article to look into how, in the context of Vietnam, implementing digital transformation necessitates the creation of a financial system. Especially in light of the devastating effects of the COVID-19 pandemic, Vietnam is giving greater attention to implementing digital transformation. Unfortunately, due to a paucity of pertinent data, researchers are unable to examine these links using conventional methods. By providing fresh statistical data on the function of finance, this study contributes to the body of information regarding the causes of digitalization. Second, the study uses is an extended time-varying autoregression (ETVP-VAR) model, which is a more recent time series data econometric technique. This method is helpful for examining the energetic connections between digitalization, environmental issues and economic globalization in Vietnam, where data accessibility is a big problem (Ha, 2022a; Ha and Nham, 2022). This method allows us to tackle the problem of data shortage. More notably, our techniques have the advantage of helping us identify the transmission pathways between digitalization, environmental issues and economic globalization. With the help of this approach, we can address the data shortage issue. The findings of this study can aid Vietnam’s efforts to develop its digital economy.

This study uses empirical research to analyze the cause-and-effect relationship between digital uptake, globalization, and pollution emissions. Our research finds that globalization is an essential but inadequate prerequisite for adopting digital technologies and that the adoption of digital technology is positively connected with levels of globalization. As a result, strategies to accelerate the adoption of digital technology in an economy should prioritize improving the globalization index, resulting in more knowledge and competitiveness spillover. It is divided into five sections: a literature review, hypotheses, data, methods, study findings, and consequences for future research.

Literature review

Theoretical foundation

Krugman (1995) and Krugman (1979) established new and endogenous growth models that stress the need for export diversification. Due to high productivity and company size variation across sectors, Krugman’s monopolistic competition trade model, which is commonly employed in modern trade theory, fails to explain actual data. Melitz (2003) presented the trading model of heterogeneous firms (HFTM) to solve this issue, which includes a self-selection process that relates trade costs to aggregate trade. Lower trade costs result in fewer export items and export decisions by enterprises. On the other hand, a rise in the number of exporters improves export diversity and boosts a country’s export growth.

Melitz’s heterogeneous firm trade model, introduced in 2003, established a link between trade expenses and total trade, resulting in increased export diversification as the number of exporters increased. However, this hypothesis did not consider the implications of trade costs on the intense margin of trade. Lawless et al. (2019) offered a novel hypothesis highlighting two ways in which trade expenses might affect overall exports: the export productivity threshold and current exporters’ sales. Highly productive enterprises may drive less productive firms out of the market, resulting in product concentration and an increase in exports at an intense margin. However, predicting changes in the intensive margin when variable expenses are included is problematic.

Trade theories highlight the importance of trade expenses used to explain export diversification. These costs include both permanent and variable charges, with the former relating to the costs of establishing distribution channels and bureaucratic processes involved with exports, such as document preparation and port procedures. Firms endure these expenses because of insufficient institutional mechanisms. As a result, improving institutional quality/efficiency can help enterprises adapt their production processes to the global market and distribute their resources more efficiently, resulting in enhanced economic diversification. The second form of trade cost, variable costs, is impacted by all exporters’ actions. On a per-unit basis, Melitz (2003) claims that these costs are of the “iceberg” sort (e.g., transportation charges, tariffs), whereas Samuelson (1952) defines variable expenses on the iceberg as a tax equivalent dependent on the value of products being exchanged.

According to the study, different facets of technological preparation can impact user behavior to differing degrees, with personality and social influence being the most powerful variables. Olschewski et al. (2018) discovered that social impact has a bigger impact on the human acceptance of technology than technological readiness. Foster and Rosenzweig investigated the variables that impact technology selection and input allocation decisions in their 2010 study. Their research goes beyond typical adoption studies, which focus on early acceptance and give insights into technology uptake and usage throughout time.

Adoption, learning, and social learning, financial and non-financial benefits, education, technology externalities, credit limits, risks, economies of scale and partial insurance, and aberrations from social standards predicted by basic rationality models were among them. The writers believed that introducing and efficiently using new technology where education is one of the most important components of the growth process.

Several studies have been conducted to investigate the link between technology adoption and entrance obstacles. According to Fang (2017), lowering entry costs to the level of the United States can result in a large improvement in productivity. Bridgman et al. (2007) established a political economy model that considered the interest groups’ authority to reject non-members and the full benefits and dispersed costs of technology adoption. According to the report, protection and interest lobbying raise technological adoption barriers, lowering worker productivity and economic growth.

Kwon and Chun (2015) explored how local enterprises strategically acquire transnational technologies and the technological transfer impacts. Local businesses may choose not to use widely used technologies in order to avoid competing with global corporations in the local market. Basly and Hammouda (2020) state that a defined model of digital entrepreneurship is required to accelerate the digital adoption process in family companies, particularly small and medium-sized organizations. Jahanmir and Cavadas (2018) discovered that consumer innovativeness, consumer attitudes toward technology, and lead-user profile all influence the late adoption of innovation. According to Lee et al. (2020), the key driving forces for digital transformation in small and medium-sized businesses are sales, marketing, process improvement, and product development.

The present literature on technology adoption does not thoroughly examine the relationship between globalization and technology adoption. According to Fatima (2017), globalization might stifle local businesses’ efforts to innovate, requiring them to innovate through product or process innovation in order to remain competitive. According to Sag et al. (2019), the key factors of open innovation are digital technology and globalization. Huggins et al. (2020) emphasize the need for university-industry collaboration for open innovation. Product innovation is more prevalent in globalization-induced innovation than process innovation.

High- and middle-income nations see increased technological and intellectual transfer impacts than low-income ones. Developing countries are not enjoying the maximum advantages of IT technology adoption. Ejiaku (2014) investigated the barriers to the adoption of information technology in poor countries. Inadequate infrastructure, government policies, and insufficient training and credentials all contribute to difficulties in transferring and utilizing information technology.

Meschi et al. (2011) explored the link between technological adoption, openness, and the relative scarcity of skilled workers in Turkey’s industrial industry. The introduction of technology raises the demand for higher-skilled human resources. Many studies make use of trade, FDI, and openness data. However, there are few studies that use direct indicators to investigate globalization and its impact on digital technology uptake. In this study, we attempted to fill a vacuum in the globalization literature. We saw it directly using the (KOF) index and levels of technology adoption (Gygli et al., 2019; Potrafke, 2015).

Digitalization and globalization performance

Based on earlier theoretical models, notably the heterogeneous trade model, we want to establish a link between digitalization and globalization or international business activities. These models imply that the process of digital transformation might involve both fixed and variable globalization or costs for international business activities. The prior study suggests that The Internet has the ability to improve commerce by decreasing fixed expenses relevant to the market and giving enterprises a competitive edge. For example, Freund and Weinhold (2004) show that the Internet lowers market-specific fixed costs of trade and increases the competitive benefits companies obtain from using the Internet. Similarly, Bojnec and Ferto (2010) discovered that the Internet has a favorable and considerable influence on food exports, owing mostly to its influence on market-specific entry costs for the food business, which is variable trade costs.

Moreover, digitalization substantially influences the international business environment, giving enterprises a competitive advantage and enabling internationalization (Lee and Falahat, 2019). In uncertain times, digitization might assist in reducing negative impacts on exports by boosting access to products and services (OECD, 2020), expediting cross-border goods movements, and increasing openness and accessibility in commercial procedures. The adoption of digital technology in manufacturing and commercial operations has increased productivity and the efficiency of communication between clients and providers (Rehnberg and Ponte, 2018).

The fourth industrial revolution has expanded the digital economy, which has resulted in the development of innovative business concepts and important resources (Bettiol et al., 2021). In the manufacturing industry, several technologies like big data, robots, the Internet, artificial intelligence, and component manufacturing have been widely used. As a result, companies have steadily embraced technology to improve the efficiency of their production and commercial processes. This technological approach has revolutionized the organization and operation of international trade (Alcácer et al., 2016).

According to Laplume et al. (2016), using new technologies in the manufacturing process via digital transformation can disrupt the global specialization trend by separating production into numerous pieces that are closer to customers. Large multinational corporations such as Amazon and Alibaba are exploiting digital platforms to give consumers quick and efficient goods delivery. Firms seeking to access global markets more successfully embrace digital transformation to reach a particular competitive level (Strange and Zucchella, 2017).

Exporting enterprises have been able to efficiently and comprehensively advertise their products in international markets as technology and advertising have advanced. This has resulted in firms extending their operations in order to collect more data, but it also has complicated product distribution owing to the rigorous internationalization process and increasing rivalry.

Organizations must welcome technological innovation to boost production efficiency in order to compete in the increasingly competitive global market. Export success is dependent on technical progress since it allows enterprises to charge higher prices, operate at lower costs, or both. When compared to competitors in the same area, this leads to more revenue and faster growth. By boosting innovation and adoption of digital transformation, exporting enterprises and their countries may attain better growth (Porter and Heppelmann, 2015).

According to Dalenogare et al. (2018), digitalization is an important part of industrialization that may increase the efficiency of a company’s production process. The Internet, Big data, and artificial intelligence may improve communication and technology implementation for exporting organizations, resulting in improved performance. To successfully adapt to foreign markets, organizations should use a method of partial invention to their IT systems rather than completely revamping them, as Azar and Ciabuschi (2017) advocate.

Digitalization and diversification of globalization or international business activities

However, studies suggest that digitization has little influence on globalization. Diverse scholars argue on the transnational influence of digital technology from diverse angles. Technology, according to Rehnberg and Ponte (2018), may change how production is bound by geography and time, affecting the location and size of production activities. Strange and Zucchella (2017) analyze the possible impact of big data on exports. Companies may learn about emerging trends in other nations without the need to sell their goods in other countries.

Firms may efficiently distribute their items to a wide variety of customers in worldwide markets while also enhancing their manufacturing processes by employing technology and data analysis. Careful research and analysis may assist businesses in identifying possible growth prospects, particularly in underserved emerging areas. These markets provide a one-of-a-kind chance for firms wishing to expand, particularly if there is little rivalry from current businesses in the region.

The public is getting more concerned about the environment and cybersecurity as organizations become more linked (Sulich et al., 2021). To address sustainability problems, the environmental goods and services sector (GGS) has grown to be a significant and rapidly expanding business, notably in the European Union (Scholz, 2017; Sergi et al., 2019). The growth of the ICT industry and network entities is viewed as vital for enhancing environmental management and boosting GGS growth. Firms may greatly contribute to the expansion of the GGS by including the Sustainable Development Goals (SDGs) in their strategy, and digitalization is a crucial instrument for doing so (Sulich et al., 2021). As a result, adopting environmental technology and solid cybersecurity policies is critical for businesses seeking long-term success (Sulich et al., 2021).

Industry 4.0 has substantially enhanced automation and communication in numerous industries, according to Awan et al. (2021), by implementing self-monitoring and intelligent machines. Furthermore, when connected to the circular economy, Industry 4.0 introduces a new industrial paradigm that enables novel techniques for more sustainable resource management (Centobelli et al., 2020). Some experts, however, contend that concentrating simply on technical efficiency to attain long-term viability is inadequate (Hilty et al., 2011; Hovardas, 2016). One key problem in bringing about the digital revolution in current life is maintaining cybersecurity and biosecurity, which are required for accomplishing the SDGs (Sulich et al., 2021). To attain sustainability, European nations must prioritize internal and external security and adopt consistent policies to employ digitalization successfully (Bossong and Rhinard, 2013).

Through many pathways, digitalization has a tremendous influence on the environment. The first is that technology may improve the effectiveness of electronic trash collection and recycling or reusing obsolete materials, contributing to a circular economy. Solid trash, e-waste, food waste, and agricultural waste are all major environmental challenges that must be addressed digitally (Sharma et al., 2020; Wen et al., 2021). Positively, digital technology has the potential to reduce the strain on biodiversity and the natural environment.

By presenting biological facts and feasible digital business models, ICT may improve the efficacy of governmental policies and citizens’ knowledge of biodiversity protection. Digitalization can also result in cost savings, improved worker safety, and a reduction in resource use and degradation in ecologically hazardous industries (Roy and Singh, 2017; Y. Zhang et al., 2017). The influence of digitalization on the relationship between ecosystems and human well-being is of particular interest to researchers since it has the potential to solve problems such as air pollution, traffic congestion, and resource scarcity (Feroz et al., 2021; Ha, 2022a; Honarvar and Sami, 2019).

There is also evidence that digitization has an impact on trade performance (Bettiol et al., 2021) and the diversity of trade (Chiarvesio and Romanello, 2018). Given the significance of digitization in encouraging long-term development, we believe the digital transformation process has an impact on the GGS. Disparities in technology adoption lead to productivity fluctuations and economic growth. Researchers have extensively researched technology adoption patterns and drivers. Acemoglu et al. (2007) created a methodology to investigate the consequences of contractual violation, technology adoption, and complementary technologies. Contractual incompletion has a stronger influence on technology adoption in businesses with more complementary intermediate inputs. Dastidar (2015) investigated how two rivals in the alternative commodities market, Cournot and Bertrand, give incentives for cost-cutting technology adoption in a vertically unique business. The authors proved that the cost and quality of new technology are regulated by the equilibrium conclusion rather than the outcome of the scoring auction.

Consumer technology acceptance differs significantly from commercial technology adoption (Forman et al., 2018). The initial application of new technology by a person, company, or other organization is referred to as technology adoption. This includes new goods, services, or managerial strategies. Using examples from the adoption of information technology, the researchers emphasize the significance of costs, benefits, communication networks., and other complicated elements in new technology deployment. According to Sadik (2008), strong institutions that reduce the cost of adoption are associated with better per capita income. According to the concept, the total costs, transportation, and institutional costs connected with adopting technology are decisive factors in the industrialization of an area.

In order for technology adoption to be a factor in globalization, a country must be open to international investment, business entry, and the transfer of technology and knowledge. Straub (2009) identified several factors when researching the adoption and dissemination theory that influence people’s technological choices. Miranda et al. (2016) revealed that adopting or rejecting technology solutions is complicated and nonlinear. Fatima (2017) investigated the connection between openness and technological adoption using data from 30 industrialized and developing countries. FDI are often less likely to innovate than domestic enterprises, particularly in the field of process innovation.

Globalization can increase chances for knowledge dissemination, but its advantages are not uniform. It is still available to developed nations (Archibugi and Pietrobelli, 2003). According to Khanna and Palepu (2010), family-owned firms rely on organizational resilience to thrive. This advantage, however, does not necessarily exist in established economies, which have better institutions and greater competition. Family-owned firms are less likely to use technology than non-family-owned enterprises. The study by Son and Han (2011) adds to our knowledge of technology adoption by demonstrating how users’ propensity to utilize technology influences their behavior after adoption.

Statistical analysis and methodology

Statistical analysis

In this paper, we use a year-by-year time series of digitalization in Vietnam for five variables: Individuals using the Internet (% of the population) (INU), mobile cellular subscriptions (per 100 people) (MOBSUB), FDI, trade openness (TRADE), and CO2 emissions (CO2). The period in our study spanned from 1996 to 2022, with data on INU, MOBSUB, FDI, and TRADE collected from the World Bank database and CO2 data sourced from the Our World in Data database. After data collection, we calculated the growth rates of the above data series and used them in the empirical analysis. The reasons for our variable selection are twofold. First, although this paper concentrated on exploring a dynamic interlinkage between digitalization and economic globalization, the system that was designed to be completed should include other dimensions of the economy. The whole system reflecting the real economy should cover the economic and environmental issues. Second, there exists an interrelation between digitalization, economic globalization, and economic and environmental issues. In particular, digitalization and economic globalization have economic and environmental effects (Ha, 2022a; Ha et al., 2022; Thanh et al., 2022). On the other side, economic and environmental factors significantly determine the success of adopting digitalization and developing economic globalization (Ahmedov, 2020; Degryse, 2016; Li et al., 2021; J. Zhang et al., 2022). In the next section, we investigate the trend of these variables. Figure 1 demonstrates a pattern of these series.

Fig. 1
Fig. 1
Full size image

INU, MOBSUB, FDI, TRADE, CO2 growth.

As can be seen from Table 1, all series have positive growth rates, indicating the increase in absolute value over time. Individuals using the Internet (% of the population) (INU) shows the largest variance. Three variables, including mobile cellular subscriptions (per 100 people) (MOBSUB), trade openness (TRADE), and CO2 emissions (CO2), have platykurtic distribution in which their tails are thinner than those of normal distribution, especially MOBSUB with the smallest kurtosis. Two remaining indicators, including the number of individuals using the Internet (% of the population) (INU) and FDI, have leptokurtic distribution, meaning that the tail is fatter than those of normal distribution. Based on the ERS unit root test suggested by Elliott et al. (1996), we figured out that neither first log-difference of any series is significant at 1%, 5%, and 10%. The correlation between the first log-difference series and their squared term expected HC is weak based on the weighted portmanteau test. Hence, it significantly supports the use of the TVP – VAR model to deal with such a time-varying variance-covariance structure.

Table 1 Summary statistics.
Full size table

Empirical methodology

One of the most widely utilized econometric techniques for examining connectedness was brought forward by Diebold and Yilmaz (2012). Researchers use this approach to track contagions in a preset system. They resolve negative impacts caused by a certain economic shock. An issue with the original method is that it relies on a time-variant of connectedness in order to determine the size of the window. In this research, we use an autoregressive vector model with time-varying parameters in conjunction with a technique to extend joint connectedness to investigate the interconnections between digitalization, green technology, pollution emissions, and energy efficiency, as described by Balcilar et al. (2021).

Time-varying parameter vector autoregression

Our article considers a one-order delay in estimating the TVP-VAR model according to the Bayesian information criterion (BIC):

$$\begin{array}{cc}{y}_{t}={{\rm{N}}}_{t}{y}_{t-1}+{\psi }_{t} \; {\in }_{t}\sim N(0,{\varSigma }_{t})\end{array}$$
(1)
$$\begin{array}{cc}{vec}\left({{\rm{ N}}}_{t}\right)={vec}\left({{\rm{N}}}_{t-1}\right)+{u}_{t} \; {u}_{t} \sim N\left(0,{R}_{t}\right)\end{array}$$
(2)

where Νt and Σt are P × P dimensional matrices, whereas \({y}_{t},\,{y}_{t-1}\) and ѱt are P × 1 dimensional vectors. Rt is a P2 × P2 dimensional matrix, whereas \({vec}\left({{\rm{{\rm N}}}}_{t}\right)\) and ut are P2 × 1 dimensional vectors. This method includes all indices (Νt) changing throughout time, as well as the connection between series. Moreover, the Σt and Rt The variance-covariance matrix is considered a time-varying matrix. A changing market and risk ratio results in varying variances and covariances, particularly in the financial market.

In the following step, we transform TVP-VAR into TVP-VMA based on Wold’s representation theorem: \({y}_{t}={\sum }_{h=0}^{\infty }{N}_{h,t}{\Psi}_{t-1}\) where \({N}_{0}=\,{I}_{Z}\) and ѱt is a vector of white noise shocks (symmetric but not orthogonal) with P × P time-varying covariance matrix \(E\left({\Psi}_{t}{{\Psi}^{{\prime} }}_{t}\right)={{\Sigma }}_{t}\). As a result, the η-step estimate error is expressed as:

$${f}_{t}\left({\eta }\right)=\,{y}_{t+{\eta }}-E({y}_{t+{\eta }}|{y}_{t},\,{y}_{t-1},\ldots )$$
(3)
$$=\mathop{\sum }\limits_{l=0}^{\eta -1}{N}_{l,t}{\Psi }_{t+\eta -l}$$
(4)

with a forecast error covariance matrix equal to:

$$E\left(\left(\right.{f}_{t}\left({\eta }\right){f}^{\prime} _{t}\left({\eta }\right)\right)={N}_{l,t}{\Sigma }_{t}{N}^{\prime} _{h,t}$$
(5)

The suggested approach is based on Koop et al. (1996). η is a significant step forward for the decomposition of forecast error variance in general (GFEVD The (scaled) GFEVD \(q{d{\varepsilon }}_{{ij},t}\), describes the impact of a shock in indicator j on indicator i and is written as follows:

$${f}_{{ij},t}^{{gen}}({\eta })=\frac{E\left({f}_{i,t}^{2}\left({\eta }\right)\right)-E{\left[{f}_{i,t}\left({\eta }\right)-E\left({f}_{i,t}\left({\eta }\right)\right)|{\Psi}_{j,t+1},\ldots ,{\Psi}_{j,t+1}\right]}^{2}}{E\left({f}_{{it}}^{2}\left({\eta }\right)\right)}$$
(6)
$$=\frac{{\sum }_{l=0}^{{\eta }-1}{\left({{e}^{{\prime} }}_{i}{N}_{{lt}}{\Sigma }_{t}{e}_{j}\right)}^{2}}{\left({{e}^{{\prime} }}_{j}{\Sigma }_{t}{e}_{j}\right).{\sum }_{l=0}^{{\eta }-1}\left({{e}^{{\prime} }}_{i}{N}_{{lt}}{\Sigma }_{t}{{N}^{{\prime} }}_{{lt}}{e}_{i}\right)}$$
(7)
$$q{d{\varepsilon }}_{{ij},t}=\frac{{f}_{{ij},t}^{{gen}}({\eta })}{{\sum }_{j=1}^{{\eta }}{f}_{{ij},t}^{{gen}}({\eta })}$$
(8)

where ei is a P × 1 zero selection vector with unity on its ith location and \({f}_{{ij},t}^{{gen}}({\eta })\) is the decreased level of indicator i’s η-step prediction error variance owing to controlling the unexpected shocks of indicator j.

Numerous spillover summary estimates are derived from the generalized spillover table, including the connectivity in all directions from indicator i to others. The following is an expression for this statistic:

$${B}_{i\,\leftarrow {{\bullet }},\,t}^{{gen},{from}}=\,\mathop{\sum }\limits_{j=1,\,i\ne j}^{{\rm{P}}}q{d{\varepsilon }}_{{ij},t}$$
(9)
$${B}_{i\,\to {{\bullet }},\,t}^{{gen},{to}}=\,\mathop{\sum }\limits_{j=1,\,i\ne j}^{{\rm{P}}}q{d{\varepsilon }}_{{ij},t}$$
(10)

The connectivity in all directions of indicator I is a key metric as part of the connectedness approach, which indicates whether indicator I impact the system more than it is affected by it: \({B}_{i,{t}}^{{gen},{net}}=\,{B}_{{i}\to \,{\bullet },{t}}^{{gen},{to}}-\,{B}_{{i}\leftarrow \,{\bullet },{t}}^{{gen},{from}}\). If \({B}_{i,{t}}^{{gen},{net}} \,>\, 0\left({B}_{i,{t}}^{{gen},{net}} \,<\, 0\right)\), the indicator i is a net receiver (transmitter) of shocks; that is, it is driving (driven by) the system.

An important component of the connectedness center is the total connectedness index (TCI), which displays the interconnections within a system. The directional connectivity is calculated as follows:

$$q{d{\varepsilon }}_{t}=\,\frac{1}{{\rm{P}}}\,\mathop{\sum }\limits_{i\,=1}^{{\rm{P}}}{B}_{i\,\leftarrow {{\bullet }},\,t}^{{gen},{from}}=\,\frac{1}{{\rm{P}}}\,\mathop{\sum }\limits_{i\,=1}^{{\rm{P}}}{B}_{i\,\to {{\bullet }},\,t}^{{gen},{to}}$$
(11)

As such, a large value implies a large level of system spillover and a large amount of market risk, while a small value implies a small level of system risk. From a portfolio diverseness perspective, there is no effect on other indicators from shocks in one indicator.

Furthermore, net pairwise directional spillovers are a way to demonstrate the pairwise interactions between two indicators through the connectedness method: \({B}_{i,{t}}^{{gen},{net}}=\,{{qd}{\varepsilon }}_{{ij},{t}}^{{gen},{to}}-\,{{qd}{\varepsilon }}_{{ij},{t}}^{{gen},{from}}\). If \({B}_{{ij},{t}}^{{gen},{net}} \,>\, 0\left({B}_{{ij},{t}}^{{gen},{net}} \,<\, 0\right)\), the indicator i dominates the indicator j, it has a greater effect on the indicator j than the indicator j has on indicator i.

Extended joint connectedness approach

The study aims to assess the \({{qd}{\varepsilon }}_{{ij},t}\) a method of mutual connectedness that is equivalent, called \({jd{\varepsilon }}_{{ij},t}\), these criteria are met by an organization:

$${B}_{i\,\leftarrow {{\bullet }},\,t}^{{jnt},{from}}=\,\mathop{\sum }\limits_{j=1,\,i\ne j}^{{\rm{P}}}{{jd}{\varepsilon }}_{{ij},t}$$
(12)
$${B}_{{{\bullet }}\leftarrow \,i,\,t}^{{jnt},{to}}=\,\mathop{\sum }\limits_{j=1,\,i\ne j}^{{\rm{P}}}{{jd}{\varepsilon }}_{{ji},t}$$
(13)
$${{jd}{\varepsilon }}_{i}=\frac{1}{z}\mathop{\sum }\limits_{i=1}^{{\rm{P}}}{B}_{i\,\leftarrow {{\bullet }},\,t}^{{jnt},{from}}=\,\frac{1}{{\rm{P}}}\mathop{\sum }\limits_{i=1}^{{\rm{P}}}{B}_{i\,\to {{\bullet }}}^{{jnt},{to}}$$

In order to achieve this, we have to adopt the approach of Lastrapes and Wiesen (2021). As a result, Eq. 12’s recommended computation must be correct. The table must also be the same in the diagonal components of the joint connectedness since the row totals must equal 1. Therefore, scaling factors vary per row, resulting in the following formula:

$$\,{\eta }_{i}=\,\frac{{B}_{i\,\leftarrow {{\bullet }},\,t}^{{jnt},{from}}}{{B}_{i\,\leftarrow {{\bullet }},\,t}^{{gen},{from}}}$$
(14)
$$\eta =\,\frac{1}{{\rm{P}}}\mathop{\sum }\limits_{i=1}^{{\rm{P}}}{\eta }_{i}$$
(15)

Due to the soaring elements in each row, our soaring differs from the one that results based on the joint connectedness method. As a result, the following steps must be followed:

  • \({jd{\varepsilon }}_{{ij},t}=\,{\eta }_{i}{qd{\varepsilon }}_{{ij},t}\)

  • \({jd{\varepsilon }}_{{ii},t}=1-\,{B}_{{i}\leftarrow \,{\bullet },{t}}^{{jnt},{from}}\)

  • \({B}_{{i}\to \,{\bullet },t}^{{jnt},{to}}=\,\mathop{\sum }\limits_{j=1,{j}\ne i}^{{\rm{P}}}{jd{\varepsilon }}_{{ij},t}\)

It is also possible to calculate the metrics for directional connectedness based on net totals and pairwise distances on the soaring parameter for each row:

$$\,{B}_{i,\,t}^{{jnt},{net}}=\,{B}_{i\,\to {{\bullet }},\,t}^{{jnt},{to}}-\,{B}_{i\,\leftarrow {{\bullet }},\,t}^{{jnt},{from}}$$
(16)
$${B}_{{ij},\,t}^{{jnt},{net}}=\,{{qd}{\varepsilon }}_{{ji},t}-\,{{qd}{\varepsilon }}_{{ij},t}$$
(17)

Our approach is expected to provide more precise findings than methods employed in previous studies because there is a similar explanation for the original connectivity methods proposed by Caloia et al. (2019). In general, this approach addresses several concerns associated with previously proposed connectedness strategies, such as: (i) it is not necessary to choose a rolling size at random; (ii) utilizing the Kalman gain of multivariate Kalman filtering, the predicted results are not influenced by outliers; (iii) The VAR coefficients can fluctuate over time; and (iv) Portfolio and risk administrators should take note of this to be able to observe the volatility of the energy market over time by allowing variances and covariances to fluctuate; (v) the solution of Lastrapes and Wiesen (2021) has been implemented a normalization problem involving row sums; and (iv) is similar to the directional joint connectedness study, the connectivity of joints is enlarged in a special way, but it enables greater flexibility and multiple directional connectedness measures to be computed. This strategy emphasizes the comparative bilateral power of the indicators, which is one of its core features.

Results

Baseline findings

Table 2 shows the average result of the interconnectedness of indicators within the network. The in-diagonal part shows the variation of a single indicator estimated by its own shocks. Off-diagonal parts report how this single indicator affects the fluctuation of other indicators in the network (FROM) and how other indicators within the network affect the fluctuation of a single indicator (TO). Estimated results on the rows indicate the influence of each indicator on the predicted error variance of a specific indicator, while the columns show the impact of one indicator on all other indicators separately.

Table 2 Averaged joint connectedness.
Full size table

The average value of TCI is 40.26%, meaning that within-network fluctuation accounts for 40.26% of the variation in the network of indicators, or so-called idiosyncratic impact of the network of variables. The last row in Table 2 summarizes the role of each variable in the variation of the network. INU plays the major role of net shock transmitter in the network, representing 15.57%. FDI transmits marginally net shocks, only accounting for 0.34%. The three remaining variables are net shock receivers, in which TRADE absorbs the highest share of net shocks (−10.51%), followed by CO2 at −4.10%. MOBSUB accounts for the smallest share of net shock receivers, only at −1.30%.

In this part, we examine more closely the dynamic behavior of variables when changing from net shock receivers to net shock transmitters over time and vice versa. We begin by investigating the time-variant approximations of total connectedness. In Fig. 2, the extended joint connectedness approach is illustrated by the black-shaded area, while the red line depicts the joint connectedness from the TVP – VAR approach. It stands out that the two approaches show a relatively similar magnitude of connectedness over time, in which the trend, peaks, and troughs are relatively relevant. The difference between the two approaches arises from the number of connections across the sample period. TCI value increased from 36% at the beginning of 1996 to 45% in 1999, then remained relatively stable until 2006. It can be attributed to the outbreak of internet users in Vietnam when telecommunication and the Internet were preliminarily set up. More specifically, from 2000 to 2003, the number of internet users in Vietnam rose sharply from 800,000 to 2 million. After that, TCI fell to a low in 2007 before recovering to around 40% in 2009. This level was stable until the end of the investigation period.

Fig. 2
Fig. 2
Full size image

Dynamic total connectedness.

Regarding the relationship between digitalization and globalization, the literature has also indicated similar findings. Particularly, as a result of globalization, technology adoption is closely linked to the country’s openness to foreign investment, the number of foreign firms entering the country, the level of technology adoption, and the transfer of knowledge. According to Straub (2009), various factors influence a person’s decision to adopt a technology. Based on Miranda et al.’s (2016) findings, specific technological solutions or innovations are adopted or rejected in a diffuse and nonlinear manner. An investigation of the relationship between openness and technology adoption was conducted by Fatima (2017) using data from 30 developed and emerging economies. FDI is less likely than its domestic counterparts to innovate in terms of process innovation. There has been an increase in the dissemination of information as a result of globalization, but this does not guarantee that all nations and organizations will be able to benefit from it. Archibugi and Pietrobelli (2003) do not explicitly exclude developing countries. Several factors contribute to the sustainability of family-owned businesses in this environment, according to Khanna and Palepu (2010). In general, the empirical evidence of this nexus in developing economies is quite limited. Our study provides empirical evidence on the interlinkages between digitalization and globalization in the case of a developing country. In the literature, a lack of relevant data in developing countries hinders scholars from investigating these relationships.

Only a few studies have applied a similar approach to the one we have used in this paper. It has been found in the literature that ecological sustainability, as well as the growth of green energy consumption, dominate all other markets and are the major shock transmitters in the entire network (Attarzadeh and Balcilar, 2022). As a result of network shocks, globalization has been the net recipient. Consequently, if the openness level of a country reaches a certain level, it will play an important role in absorbing the dynamic volatility in the energy market. In the energy market and the environment, innovation or technological progress plays a similar role in minimizing risks. According to the results, innovation, economic development, and globalization can help a country by mitigating dynamic volatility in the energy market by absorbing it or transferring it to other markets. We provide evidence of globalization and digitalization in a developing economy, contributing to the literature.

To better understand the dynamic connectedness over time of the system, we zoom in on the behavior of each indicator. Figures 3–5 illustrates the interchangeableness of the net shock receiver and net shock transmitter’s status of variables over time. A positive return means that the indicator transmits into the network of indicators, while a negative result indicates that the variable absorbs the shock of the network. As can be seen from Fig. 3, INU is the major net shock transmitter to the system. The highest effect of shock transmitters by INU was recorded in 1999, reaching 20%. In the following year, the transmission of net shocks was less intensive, decreasing slightly lower than 20%. This behavior can be explained by the outbreak of Internet users in Vietnam, which started in 1997, in which internet users surged from 200,000 to 3 million in 2002 and 20 million in 2007. Since then, the Internet penetrated Vietnam significantly, reaching relatively stable from year to year around 50–70%. TRADE functioned as a net shock transmitter during a very short beginning period, then turned into the net shock receiver till the end of the period. The role of the net shock receiver of trade openness has become increasingly significant since 1996, reaching a peak in 2005 before declining nearly to 0% in 2007. Before 2007, the connectedness of trade openness was significantly high and unstable because the economy was in transition. After joining WTO in 2007, the connectedness of trade openness absorbed around 20% of net shocks for the network of indicators. Meanwhile, MOBSUB and FDI acted as net shock absorbers from 1996 to 1999, then switched to net shock transmitters from 1999 to 2007. FDI has been reversed to a net shock receiver since 2007. The FDI inflow in Vietnam was significantly high after joining in 2007. FDI plays an important role in improving infrastructure, especially the Internet, and establishing a connection to the global network.

Fig. 3
Fig. 3
Full size image

Dynamic net total directional connectedness.

Fig. 4
Fig. 4
Full size image

Dynamic net pairwise directional connectedness: other variables’ volatility to INU.

Fig. 5
Fig. 5
Full size image

Dynamic net pairwise directional connectedness: Changes in INU to other variables’ volatility.

A global economic downturn occurred between 2006 and 2008. This event affected all aspects of the economy, not only the innovation and technology sectors. Because the government took taxpayers’ money to bail out failing financial institutions, the public has become distrustful of banks and is seeking alternatives. Several FinTech companies have arisen as a result of the younger generation’s native use of technology. There is no doubt that the warmer region is expected to expand even further in 2014 and beyond. Digital economies and their ongoing processes have a significant impact on a country’s competition in the global market. Vietnam, however, appears to be experiencing a stagnation or slow implementation of changes in this area. As a result, their competitive ability in the global market will be required to improve, which could pose a problem in the future. As a result of these trends, globalization matters, and the possibilities for developing them need to be given more attention.

In the remaining period, two indicators changed to net shock absorbers in the networks, in which their capacity to absorb net shocks was relatively the same. CO2 emissions behaved oppositely in terms of net shock absorber and transmitter compared to the two above variables. This variable transmitted net shocks to the system from 1996 to 1998, then absorbed significant net shocks to the system (up to 40%). The role of the net shock transmitter from 2007 to 2019 of CO2 was relatively modest, around 10% of the total net shocks of the system. There has been a 50% increase in 3G traffic on Vietnamese networks since the end of 2013 and a significant increase in the number of 3G subscribers. Further, mobile cellular subscriptions and carbon dioxide emissions have been characterized as playing a similar role, with the latter acting primarily as a net receiver of shocks. The Vietnam Manufacturing Expo 2013 was one of many events that Vietnam held during that period to attract foreign investments in technology and innovation. In some cases, companies or businesses may invest in new technology. As an example, FPT has decided to “bet” on three of the most current technologies - cloud computing, mobility, and big data - for its new development phase.

Furthermore, Fig. 6 displays the network connectivity metrics in 1996 and 2005. In 1996, trade openness dominated FDI, CO2 emissions, and mobile users, while internet users were dominated by FDI, mobile users, and CO2 emissions. The domination of FDI increased in 2005 and played an important role in developing digitalization in Vietnam.

Fig. 6
Fig. 6
Full size image

Network connectedness.

Other variables, except for CO2, consistently play the role of net shock receivers from INU, in which FDI absorbs net shocks considerably from INU, over 10%. The role of net shock receivers from INU of MOBSUB increased substantially from 5% in 1996 to 10% in 2019. The role of trade openness in receiving shocks from INU was marginal. CO2 emissions primarily transmitted net shocks to individuals using the Internet (% of the population), except for a short period from mid-1999 to mid-2001 in which CO2 emission absorbed shocks from INU. There are several reasons for such a phenomenon. After the economic reformation, the manufacturing sector experienced a sharp increase in Vietnam. To improve infrastructure quality, the construction of roads and buildings. Furthermore, the majority of industrial activities in Vietnam have been high emitters, triggering higher CO2 emissions. What’s more, increasing economic activities also cause rising transportation, which is a large source of CO2 emissions.

Further findings from the simulated quarterly database

To confirm our findings, we use the Automatic Structural Time Series Model (ASTSM), which is designed to automatically detect the appropriate decomposition for a univariate time series into trend, seasonal, and cycle components using a state space approach. The package also has the functionality to perform structural interpolation, i.e., interpolate a quarterly series into a monthly series using either end-of-period, period average, or period sum methods. The unobserved components are estimated with the Kalman filter and all model parameters are estimated using maximum likelihood. The Kalman filter and smoother are implemented, so they are reasonably fast. Our paper applies a time-varying parameter vector autoregression (TVP-VAR) in combination with an extended joint connectedness approach to prove interlinkages between digitalization volatility and FDI from Q1_1996 to Q1_2022. Our findings reveal that Internet users were net shock receivers especially before 2008 and from 2015 to 2018. FDI turned into a net shock receiver from 2007 onwards. The domination of FDI increased in Q4_2005 and played an important role in developing digitalization in Vietnam. This significant relationship highlights the critical role of designing FDI policies to promote more comprehensive digitalization processes.

Table 2 shows average findings for interlinkages of different indicators inside the network. The diagonal part in this table reports the fluctuation of a single indicator, which is estimated by its shocks. The off-diagonal components describe the impact of this indicator on others’ fluctuation (FROM) and others on this indicator’s fluctuation (TO). Specifically, in Table 3, the rows indicate each indicator’s impact on a given indicator’s prediction error variance. However, the columns represent the independent influence that one specific kind of indicator has on all other indicators.

Table 3 Averaged joint connectedness.
Full size table

The TCI average value is 40.67%. It is shown that changes within this network can explain 40.67% of the variation in our network. This implies that idiosyncratic influences account for almost 60% of the system’s error variation. The contribution of each indicator is indicated in the last row of Table 3. This analysis implies that FDI has a significant role in transferring shock effects and volatility to other indicators within the system. Internet users are net shock receivers. Moreover, FDI transmits the most shocks to internet and mobile users. CO2 emissions, mobile users, and trade openness are net transmitters of shocks.

It is worth noting that those average findings are most relevant as an overview of the underlying interconnectedness. In the wave of the COVID-19 epidemic and the Ukraine-Russia Conflict, average findings are restricted to examining interconnectedness across a network of factors. As a result, a more dynamic framework of analysis is required. It not only considers the change of the TCI over time but also how the function of certain indicators within the research network may vary over time.

Our study begins with dynamic total connectedness results, which show the TCI’s intertemporal volatility in Fig. 7. The TCI values vary significantly during our sample period. They reach a peak of around 50% in 2005. The larger the TCI values, the greater the level of connectedness between the indicators. It is remarkable that the TCI values tend to be stable at around 35% during 2009–2022.

Fig. 7
Fig. 7
Full size image

Dynamic total connectedness.

Our article uses net total connectedness. The positive and negative values show the net transmitting and receiving roles of each indicator under examination, as shown in Fig. 8. Internet users were net shock receivers especially before 2008 and from 2015 to 2018. FDI turned into a net shock receiver from 2007 onwards, while trade openness played an opposing role. It is worth noting that FDI played a significant transmitting role during 1996–2007. Mobile users are net shock transmitters before 2004 and from 2014 to 2018.

Fig. 8
Fig. 8
Full size image

Dynamic net total directional connectedness.

Figure 9 shows the net pairwise connectedness results. Our objective is to determine the key role of digitalization and FDI. Internet users were consistently dominated by mobile users, especially from 2015 to 2018. Similarly, Internet users were most significantly dominated by FDI between 1996 and 2007, when the gap was widest. While this dominance lessened slightly after 2007, FDI continues to control digitalization. Internet users have been dominated by trade openness and CO2 emissions from 2007 onwards.

Fig. 9
Fig. 9
Full size image

Dynamic net pairwise directional connectedness.

Furthermore, Fig. 10 displays the network connectivity metrics in Q1 1996 and Q4 2005. In Q1 1996, mobile users dominated Internet users, FDI, and CO2 emissions, while internet users were dominated by FDI, mobile users, trade openness, and CO2 emissions. The domination of FDI increased in Q4_2005 and played an important role in developing digitalization in Vietnam.

Fig. 10
Fig. 10
Full size image

Network connectedness in Q1 1996 and Q4 2005.

Conclusions and policy implications

Our study estimates the interlinkages of five variables proxied for digitalization and globalization, i.e., internet user, trade openness, FDI, mobile cellular subscription, and CO2 emission, using the dynamic connectedness approach. Yearly data from 1996 to 2019 from Vietnam are extracted from the World Bank database and Our World in Database. The TCI value for the whole sample varied between 36% and 45%, indicating that all variables are moderately interconnected. Our study shows how each variable’s role in our designed system changes over time by analyzing the time-variant of net totals and pairwise directional connectedness. Internet users consistently transmit shocks to the system, while trade openness absorbs shocks from the system throughout the period. Dynamic pairwise connectedness investigates how one variable can transmit to/receive from the other variable. Our results suggest that shocks from internet users are mainly absorbed by mobile cellular subscriptions and FDI while it receives shocks from CO2 emission since 2002. In general, we can conclude that the network of variables is subject to moderate market risks.

In a transition economy such as Vietnam, this study was designed to propose the theoretical interlink between environmental sustainability and digital transformation in the context of globalization. The literature has provided evidence on the role of digitalization in promoting economic growth, mitigating pollution emissions, or enhancing the efficiency of the financial market and government policies. On the one side, globalization is a key enabler of digital transformation within an organization. On the other hand, we also realize that the implementation of the digital transformation process helps promote globalization. Pollution mitigation, economic globalization and digital transformation have been examined in research results.

Our findings have important implications on the policy front for investors and authorities, as well as demonstrate the contagions across variables and their interrelationships. Understanding the key antecedents of these contagions allows policymakers to design the most effective policies to reduce these variables’ vulnerabilities and minimize the spread of uncertainty or risk among them. It is evident from our findings that unexpected and highly uncertain events have become increasingly interconnected. We also highlight the impacts of one variable on other variables: how shocks are transmitted or received. Further, the conclusions of this study can also be useful for policy in an effort to enhance Internet and ICT adoption, which can be attributed to the direct relationship between Internet users and trade openness, FDI, CO2 emissions, and mobile cellular subscription. It calls for policy design targeting improving the country’s digital adoption to maximize benefits from globalization.