Innovational duality and sustainable development: finding optima amidst socio-ecological policy trade-off in post-COVID-19 era

2.1 Impact of technological innovation on sustainable development goals and carbon dioxide emissions

In recent times, there is a rising interest among scholars to explore the effect of technological innovation on SDGs. Table 1 contains a summary of some of the major studies in the domain. While Mensah et al. (2018) explain how the technological development of the Organization for Economic Co-Operation and Development (OECD) countries influences their CO₂ emissions over 1990–2014, Modgil et al. (2020) propose a modern information decision support system for achieving the SDGs. Sinha et al. (2020) investigate the effect of technological innovations on environmental degradation, sustained economic growth, clean and affordable energy, and quality of education for the period 1990–2017. Bag et al. (2021) opine that big data analytics-based artificial intelligence can be an important factor for the operations of circular economy and achieving the SDGs. Chien et al. (2021) propose an information and communication technology (ICT)-based framework for the BRICS countries (Brazil, Russia, India, China, and South Africa) to combat environmental degradation and achieve the SDGs.

The exploration of the existing literature related to SDGs highlight the researchers focus on diverse topics such as performance indicators of food loss reduction (Al-Dalaeen et al., 2021), the influence of political leadership on the SDGs (Grover et al., 2021), information decision system-based framework design (Modgil et al., 2020), the impact of technological innovation on achieving SDGs (Sinha et al., 2020), humanitarian decision-making (Marić et al., 2021), etc. However, the investigation into the effect of technological innovation attaining the SDGs and on the social indicator such as unemployment has not been paid enough attention. This is the gap our study aims to fill.

2.2 Impact of technological innovation on employment

Many researchers investigate the impact of technological progress on the employment of the economy. As per the analysis, there exist two contradictory opinions. A group of researchers opines that technological innovation raises the employment level of a nation. For instance, Greenan and Guellec (2000) investigate the impact of technological innovation on the employment opportunities for the French firms over 1986–1990 and find the strong positive effect of innovation at the firm level as well as sectoral level. Similarly, Benavente and Lauterbach (2008) conclude that the technological innovation policy of Chile generated new jobs over 1998–2011. Another stream of research that has recently been of much interest to scholars studying industrial revolution and employment implications is “Industry 4.0” (Koh et al., 2019; Xu et al., 2018). Industry 4.0 is also expected to significantly enhance firm productivity and employment in coming years (Ortt et al., 2020; Opazo-Basáez et al., 2021).

Another group of researchers believes that technological progress affects employment opportunities. Evangelista and Savona (2002) conduct an empirical study on the Italian firms over 1993–1995 and exhibit the negative effect of innovation on employment. Similarly, Gagliardi (2019) investigates the effect of foreign technological innovation on domestic employment through the industry mix of the UK over 2000–2007 and finds the reduction in employment by around 5%. Recently, Zhu et al. (2021) present a dual effect of technological innovation in case of Chinese firms, i.e. process and product innovation shows the positive and negative impact, respectively, on employment opportunities. These studies indicate the need to be cognizant of technological innovations’ impact. Higher unemployment might lead to social unrest, even though it is driven by noble goals of more sustainable development.

2.3 Impact of COVID-19 pandemic on sustainable development goals (SDGs)

The impact of the ongoing COVID-19 pandemic on attaining SDGs has attracted the attention of researchers. There is a rising interest in analyzing how innovations can help nations and firms meet their SDGs. Thornton (2020) opines that the pandemic has a severe negative impact on attaining the SDGs. Pan and Zhang (2020) highlight the importance of in-depth investigation into the information systems to attain the SDGs in the post-COVID era. Modgil et al. (2021) identify the role of AI in meeting firms' resilience in post-COVID-19 world.

However, our literature review finds lack of works that focus on the impact of technological innovation on achieving sustainable development goals and employment opportunities. Specially, it emerges as a matter of concern for the nations in the post-COVID-19 era. In this context, an investigation into the dual socio-environmental effect of technological innovation for multiple countries over a longer period can facilitate efficient decision-making of policymakers. It acts as the main motivation of the study. To the best of knowledge, this is the first work that investigates the effect of technological innovations on SDGs and employment in the COVID-19 pandemic era. The summary of the relevant literature and contribution of this work is presented in Table 1. It shows the gap that this study aims to fill as the first study with dataset from 100 countries analyzing impact of innovation on sustainability and employment. None of the other studies in this field relies on the dataset of this size, and they also do not attempt to balance the impact of innovation on sustainability and employment.

3.1 Problem description

As the resistance to shift from the existing energy sources to renewable ones creates deterrence into the transition, environmental quality might not start improving immediately with the rise in technological innovation. Therefore, during the initial phases of innovation, environmental degradation might not start getting reduced. However, gradual diffusion of innovation and achieving economies of scale gradually bring down the cost of these solutions, leading to their rising acceptability. This can bring forth a steady decline in the growth of environmental degradation.

At the same time, the development and deployment of these solutions require skilled and unskilled labor, which eventually creates employment in the economy. Being capital-intensive, once these solutions are deployed, they start replacing human labor. Hence, the existing systems can be more efficient. Following Arrow et al. (1961), this replacement of labor with technology (capital) creates unemployment within the economy.

This dual impact of technological innovation might be catalyzed by the structural transformation of the economy, i.e. more service-orientation of the sectoral activities might increase unemployment, while environmental degradation might fall. On the contrary, manufacturing-oriented sectoral transformation might experience an opposite effect. Moreover, energy usage patterns might also have a moderating impact on the impact duality of innovation. As the energy-intensive production practices are more inclined toward being stemmed from the manufacturing sector, higher usage of fossil fuel might have coexistence with higher employment. Moreover, the channel of international trade is also utilized to transfer the old and dirtier technologies to comparatively poorer and less developed countries for controlling the carbon footprint. Also, the transfer of greener technologies to host countries might boost innovation capabilities, which might have a consequence on environmental quality and unemployment. This impact of the globalization channel can be explained by the “Pollution Haven hypothesis” (Levinson and Taylor, 2008) and the “Pollution Halo hypothesis” (Antweiler et al., 2001). Lastly, the developmental trajectory of a nation needs to account for the balance between demand and supply of labor. Here, population growth can emerge as a crucial factor. If the population growth is higher than the demand of labor, the country is likely to suffer from the unemployment problem. Moreover, this incidence of unemployment might lead to income inequality and deterioration in living standards. This might add to the environmental degradation arising out of the unsustainable energy usage pattern.

3.2 Empirical model

Assessing the impact of technological innovation on environmental quality and unemployment entails considering a theoretical framework that can capture the evolutionary impact of technological innovation across a group of countries that may represent a wide set of innovation and enterprise-level beliefs. Hence, the present study embarks on the EKC hypothesis framework. According to the seminal work of Grossman and Kruger (1991), environmental degradation starts rising during the early phase of economic growth. Once this economic growth reaches a threshold, improvement in the living standard raises environmental awareness among citizens. Henceforth, further growth in the economy leads to a decline in the environmental degradation. Quadratic specification of this hypothesis results in an inverted U-shaped association between environmental degradation and (drivers of) economic growth. This hypothesis is capable of revealing the nonlinear evolutionary impact of policy instruments on a target policy parameter for a pool of countries (Shahbaz and Sinha, 2019). The empirical schema of this association can be represented as follows:

• X = Matrix of other explanatory variables;

• i = Sample countries; and

• t = Sample years, β0,β1,β2≠0.

Now, first-order differentiation of equation(1) with respect to TECH. and corresponding first-order condition (FOC) determine the optimum point of the association, described in equation (2):

Now, given the two conflicting policy agendas, the shape of this association can take two different forms, based on the two dependent variables. This association is expected to follow a generally accepted inverted U-shaped form for CO₂ emissions. Therefore, in this case, β1>0 and β2<0. On the other hand, the shape is expected to be U-shaped for unemployment. Hence, in this case, β1<0 and β2>0. For both the cases, the turnaround points of the association can be expressed as −β12β2 presented in equation (2). However, interpretation of this point is different for both the cases. It denotes the maximum of and the minimum of the association for CO₂ emissions and unemployment, respectively. This can be shown by the second-order differentiation of equation (1) with respect to TECH.

Now, in empirical pursuit, equation(1) can be explained as per the following:

The description of the dependent and independent variables presented in equations (3) and (4) are expressed as follows:

1. CE = CO₂ emissions;

2. UE = Unemployment;

3. TFP = Total factor productivity;

4. GLOB = Globalization;

5. POP = Population; and

6. STR = Structural transformation of economy.

As the investment in research and development might not be realized fully in the innovation output, the output indicator of innovation can be helpful to capture the effects of innovation. For this reason, in this study, TFP, an output indicator of innovation, is selected. STR is determined by the Lilien Index (Lilien, 1982), which measures the changes in labor share across primary, secondary and tertiary sectors. It can be represented as follows:

• SE: Employment share in a particular sector; and

• s=1, 2, 3: primary, secondary and tertiary sector, respectively.

3.3 Methods

3.4 Data

The study is conducted for 100 countries over 1991–2019. These countries are segregated into four different income categories (i.e. low, lower-middle, upper-middle and high), following the categorization provided by the World Bank (Serajuddin and Hamadeh, 2020). To get the detailed description, refer to Table A1 in Appendix. The data for TFP have been obtained from Penn World Table 10.0. The data for CO₂ emissions (in metric tons per capita), population (total), unemployment (% of the total labor force), energy use (kg of oil equivalent per capita), the labor force (total) and the number of labors in three sectors (agricultural, industrial and service) have been collected from the World Development Indicators (World Bank, 2021). The data for globalization have been taken from the KOF database (Gygli et al., 2019). Except STR (structural transformation of economy), all the variables are log-transformed before analysis.

While talking about the sample selection for this study, it is worthwhile to observe that the temporal span of the study refers to the pre-COVID-19 period, whereas the objective of the study is aimed at the post-COVID-19 period. Though this selection might appear to be speculative, there are some definite reasons behind this informed choice of sample. First, because of the ongoing technological innovations, the world already started experiencing betterment in environmental quality. Slowdown in manufacturing and other anthropogenic activities due to the incidence of COVID-19 only accelerated the process. Hence, it can be said that the COVID-19 outbreak complemented the environmental benefits of technological innovations, without causing any significant transformations in the environmental impact of innovations. Second, Industry 4.0 has been gradually creating unemployment across the manufacturing firms, and literature has cited its evidence (Kovacs, 2018). The convergence in the employment rate during the Industry 4.0 regime indicates the slowdown in job creation process (Gashenko et al., 2020). This slowdown in the job creation was catalyzed by the slowdown in manufacturing and other anthropogenic activities due to the incidence of COVID-19 outbreak. Hence, the COVID-19 outbreak is said to be complemented the social detriment of technological innovations, without causing any significant transformations in the social impact of innovations. As the COVID-19 outbreak did not bring any transformations in the socio-ecological impacts of innovations, and the nature of COVID-19 only intensified those impacts, hence consideration of this particular pre-COVID-19 data to prescribe the post-COVID-19 policy framework is logical.

Lastly, it is also assumed that the impact of innovation might vary according to the level of development. Hence, focusing on a single country or a group of homogenous countries might not properly demonstrate the socio-ecological impacts of technological innovation. Moreover, the development and deployment of innovative solutions depend largely on the position of a country on a developmental trajectory. Hence, analyzing this impact for the countries across several levels of income might indicate how the socio-ecological impacts of technological innovation might change with levels of development.

4.1 Analysis of country groups

The analysis of the model outcomes starts with the baseline estimation derived from adopting the Least Squares Dummy Variable (LSDV) approach, and the model outcomes are reported in Tables 2 and 3. These outcomes are estimated for the countries of four income groups classified by the World Bank. First, the results of the long-run coefficient estimation for CO₂ emissions are discussed. The model outcomes reported in Table 2 for all the four groups of countries show that the CO₂ emissions–TFP associations resemble the generally accepted inverted-U-shaped form of EKC. Now, it is worthwhile noting that except for the case of high-income countries, the turnaround points are outside the sample range. This indicates that though the prevailing economic growth trajectory is increasing the CO₂ emissions, the growth rate of emission is decreasing, and the turnaround points will be achieved during the post-COVID period. This shows that the innovation-led economic growth trajectory being trodden by these nations is pro-environmental. Also, the turnaround point for low-income countries is the highest, followed by the ones of lower- and upper-middle-income countries. This gives a comparative scenario between the prevailing policies in these countries. With the rise in income level, policymakers strive to transform the existing policies from pro-growth to pro-development. This distinction among the economic policies in these four groups of countries is revealed through the nature of turnaround points. While all these countries depend on the innovation processes, it might be possible that the development of the innovation capabilities is not ingenious. Hence, the innovative solutions might be imported. Also, it is found that the impact of globalization on CO₂ emissions is positive, except for high-income countries. The highest impact is observed in the case of low-income countries, followed by the lower- and upper-middle-income countries. This scenario can be traced back to the “Pollution Haven hypothesis,” i.e. the high-income countries try to reduce their environmental degradation by exporting the low-cost polluting technologies to the countries with a lower level of income. For this reason, it is expected that the impact of globalization on CO₂ emissions might be negative for the high-income countries, and the study outcome is consistent with this phenomenon. Now, once these technologies are employed in the production, commercial electricity is required to run these solutions. As the fossil fuel-based energy solutions remain the predominant source of energy across the majority of the countries, it is expected that the energy usage patterns will have negative environmental consequences. The present study outcome reflects the same. However, the negative environmental impact has been observed the highest and lowest for the low- and high-income countries, respectively. This situation might have arisen due to the highest and lowest share of renewable energy solutions in the energy mix of the high- and low-income countries, respectively. Apart from the energy usage pattern in the industries, the domestic energy demand also caters to the end-energy use, leading to CO₂ emissions. Hence, the impact of population on CO₂ emissions is expected to be positive, and the study outcomes also reveal the same phenomenon. However, it is noteworthy to observe that the impact is highest in the case of the low-income countries and lowest for the upper-middle- and high-income countries. This situation can be attributed to the population growth characteristics of these nations, i.e. the low-income countries have the highest population growth, whereas it is lowest for the upper-middle- and high-income countries.

The structural transformation of the economy exhibits a positive impact on the CO₂ emissions for the low-, lower-middle- and upper-middle-income countries. On the contrary, the effect is negative in high-income countries. This finding can be traced back to the impact of economic growth patterns on environmental degradation, and thereby, substantiating the policy void existing in the low-, lower-middle- and upper-middle income countries from the perspective of achieving environmental sustainability. This is an area where policy intervention might be necessary for these countries to make progress toward achieving the Agenda 2030.

After the CO₂ emissions–economic growth association, the unemployment–economic growth association is discussed. The model outcomes reported in Table 3 for all the four groups show that the unemployment–TFP associations depict a U-shaped form. This segment of the results reveals that unemployment first shrinks with the rise in innovation-led economic growth and starts increasing after reaching a threshold. In a similar fashion with the previous case, it is observed that the turnaround point for low-income countries is the highest, followed by the ones of lower-middle-, upper-middle- and high-income countries. The comparative analysis between the prevailing policies in these countries signifies that the turnaround points remain outside the sample range for all four cases. It indicates that the economic growth trajectory trodden by these countries can perhaps give rise to the issue of unemployment during the post-COVID period. When this finding is analyzed along with the previous segment of estimation results for CO₂ emissions, a policy paradox is encountered. When the innovation-led economic growth trajectory is expected to bring environmental sustainability, the same growth trajectory compels these countries to depart from achieving social sustainability. This scenario can be analyzed further by incorporating the impacts of other explanatory variables. For low-income countries, the majority of the economic activities can be categorized as labor-intensive and manufacturing-oriented. Given the innovation capacity being comparatively less than the countries in the other three income categories, increasing economic activities raise the demand for additional labor. Now, this demand might fall at a later stage with the rise in innovation capacity. For the countries of other income categories, this situation might arise comparatively sooner, as the innovation capacity of these nations is higher than those of the countries under the low-income category. Therefore, the channel of globalization utilized by the low-income countries might help in increasing manufacturing activities. On the other hand, the countries under the other income groups might utilize this channel for boosting their innovation capacity through technology transfer. Hence, globalization might exert negative social externalities for the latter group of countries by creating unemployment, whereas boosting manufacturing activities in the low-income countries might reduce unemployment by extending the labor-intensive manufacturing activities. This situation might be traced back to the classic case of “Capital-labor substitution” demonstrated by Arrow et al. (1961). The model outcomes validate this argument. Now, the impact of energy usage patterns might substantiate this insight by the elasticity values. The coefficient of energy usage is highest for the low-income countries, whereas the impacts are comparatively lower for the other three groups of countries. It exhibits that the impact of energy use is diminishing in the countries with higher income levels, as the impact of innovation might be reflected through the energy efficiency and share of renewable energy in the energy mix of these countries. From another dimension, investigating this scenario from the perspective of the labor market, it can be opined that given ceteris paribus, a higher population might create an excess supply of labor. As a result, the unemployment situation can be aggravated. Also, the model outcomes suggest that this argument might be valid for all the four groups of countries, while the impact is the lowest for the high-income countries. When these scenarios are analyzed together, the impact can be observed in terms of the structural transformation of the economy, which functions in the intersection of product/service and labor market. In the era of Industry 4.0, the transformation of economies toward being service-oriented might open up avenues for innovations. Also, the advent of automation technologies might start replacing human labor. On the contrary, raising the traditional labor-intensive manufacturing processes shrink the innovation opportunity and eventually reduces the potential unemployment possibilities. The first half of this argument is substantiated by the model outcomes for upper-middle- and high-income countries, whereas the latter half of this argument is validated by the model outcomes for low- and lower-middle-income countries. Finally, the estimation outcomes of the model reveal that the innovation-led economic growth might lead these nations toward being socially unsustainable. As a result, the countries might experience a hindrance in attaining the objectives of Agenda 2030.

4.2 Robustness check with second-generation methods

In a globalized world, it might be possible that the countries might be connected via economic spillovers. Therefore, these associations cause the interdependency between economic growth drivers of different countries. This dependence leads to a specific estimation issue, i.e. CD. In the presence of CD, the second-generation methodological approach should be adopted. Driven by this estimation agenda, first, the possibilities of CD and slope heterogeneity are checked in the data. The test outcomes reported in Table 4 suggest the presence of CD and panel heterogeneity in data. Hence, this evidence warrants the application of the second-generation methodological approach.

The CD in the data necessitates the incorporation of second-generation panel diagnostic tests for long-run coefficient estimation. The second-generation panel unit root test is conducted to understand the order of the integration among the model parameters. The results reported in Table 5 show that the model parameters to be integrated to the first order. Based on the confirmation of order of integration among the model parameters, the confirmation of long-run association among them is validated through the second-generation panel cointegration test. The results reported in Table 6 demonstrate that the model parameters to be cointegrated, and this segment of the findings warrants the estimation of long-run coefficients.

To accommodate the CD in the data, two alternative methodological approaches, i.e. cross-section augmented distributed lag (CS-DL) and dynamic common correlated effect-augmented generalized method of moments (DCCE-GMM), have been adopted. The results of the model estimations using these two methods are provided in Tables 7 and 8. For both the models, the coefficients and corresponding turnaround points demonstrate similar patterns compared to the LSDV estimation results. These model outcomes using second-generation long-run coefficient estimation methods exhibit that the prevailing policy outcomes do not undergo any changes even in the presence of economic spillovers among the nations. Hence, it can be deduced that the model estimates derived by the LSDV method are robust, and the policy-level problems highlighted by the model outcomes are prevalent, irrespective of the economic connectedness among the nations.

4.3 Analysis of full sample

In this subsection, finally, the estimation for the entire sample is conducted using LSDV, CS-DL and DCCE-GMM methods. The estimation results are reported in Tables 9 and 10. The model outcomes reported in Table 9 signify that the CO₂ emissions–TFP association resembles the generally accepted inverted-U-shaped form of EKC for all three methods. This outcome extends the findings of Amri et al. (2019) for Tunisia and Wang et al. (2019) for China. Now, the turnaround point of this association reveals a generalized tendency of the economic growth trajectory being trodden by all the countries at large. The turnaround point within the sample range shows that the countries are largely treading on an environmentally sustainable economic growth path. Therefore, it can be assumed that during the post-COVID period, the prevailing innovation-led economic growth patterns will help the nations to achieve a lower rate of CO₂ emissions, signifying sturdy progress toward the attainment of SDG 13, i.e. climate action. This particular segment of the findings is corroborated by the impact of structural transformation of the economy on CO₂ emissions. Driven by innovation, the active economic sectors exhibit a higher inclination toward the service sector. This sectoral shift from manufacturing activities reduces the environmental pressure by reducing the CO₂ emissions in the ambient atmosphere. This outcome corroborates the finding of du Can and Price (2008). This structural transformation of the economy is complemented by the globalization pattern, which is characterized by the cross-border transfer of clean technologies. This internalization of negative environmental externality by the globalization process is substantiated by the coefficient of globalization in the model outcomes. This outcome contradicts the finding of Sharif et al. (2020) for China. However, the prevailing energy usage patterns might cause a rise in the ambient CO₂ emissions, and this scenario might be worsened by the population growth. As the global energy mix is still dominated by fossil fuel-based solutions and population growth might raise the energy demand, both these factors might add to the rise in CO₂ emissions. This argument is validated by the model outcomes. This outcome falls in the similar lines with the finding of Sinha et al. (2020) for Next 11 economies. In summary, it can be concluded that the innovation-led economic growth trajectory across the world might provide a solution to the energy security and environmental degradation problems by promoting energy efficiency and renewable energy solutions. As the world has experienced a rise in technological innovations during the COVID-19 outbreak, it might be expected that these innovations my lead the nations toward achieving environmental sustainability during the post-COVID period.

Apart from the environmental impact, the innovation-led economic growth trajectory experiences a social impact. The estimation outcomes of this impact are reported in Table 10. The model outcomes reveal that the unemployment–TFP association resembles a U-shaped association, consistent across three estimation methods. This insight indicates a socially unsustainable economic growth trajectory, as unemployment rises after a decline and reaches a certain threshold. As the economic growth might raise the level of unemployment, it might affect the social balance due to the skewed income distribution. Depending on the extent of capital and technology intensiveness of the sectoral structure in the era of Industry 4.0, the level of unemployment might vary. This outcome falls in the similar lines with the theoretical findings of Boone (2000) and empirical finding of Beynon et al. (2019) for the USA. By observing the turnaround points of this association, it can be deduced that they are very close to the top percentile of the sample, despite their presence outside the sample range. It indicates that given the economic growth trajectory prevails, the world might experience a social imbalance during the post-COVID period, leading to a potential roadblock to achieve the objectives of SDG 8, i.e. decent work and economic growth. This situation can be aggravated in the countries with high population growth due to the excess supply of labor rather than the demand. The model outcomes substantiate this claim by revealing the positive impact of population on unemployment. Along with this situation, the technology transfer via globalization channel might replace labor by technology, thus raising unemployment. The positive impact of globalization on unemployment, as revealed by the estimation outcomes, validates this argument. This outcome corroborates the finding of Gozgor (2017). However, the model outcomes also divulge that the energy usage pattern might affect unemployment by boosting labor-intensive manufacturing activities. The outcome extends the findings of Mbarek et al. (2018) for Tunisia. To complement this argument, the impact of structural transformation of the economy might be crucial. Given the impact of the structural transformation to be negative on unemployment, and given the upper-middle- and high-income countries promoting technologies in place of manual laborers, it might be possible that the surplus labor force is potentially migrating toward the countries with comparatively lower income levels. Therefore, surplus labor of one country might be catering to the demand of labor in the other countries. However, this impact cannot be a sustainable one, given the high population growth in the destination countries.

As both the impacts of the innovation-led economic growth have been discussed, along with the impacts of other relevant policy instruments, it is necessary to facilitate a holistic depiction by analyzing both scenarios together. Graphical representation of both the associations is provided in Figure 1, where the CO₂ emissions–TFP association is inverted U-shaped and the unemployment–TFP association is U-shaped. As the rise in TFP affects the CO₂ emissions and unemployment in opposite ways, the policy trade-off situation becomes evident in Figure 1. As the best solution might not be achievable in such a scenario, an optimum policy mix between these two competing objectives needs to be devised. In Figure 2, a vivid depiction of this situation is presented. It is assumed that both CO₂ emissions and unemployment are at equilibrium at point PI1. At this point, CO₂ emissions and unemployment are denoted by CE1 and UE1, respectively. This equilibrium is achieved during the pre-COVID situation. Beyond this point, both CO₂ emissions and unemployment start exhibiting a decline in their growth rates, and both these target policy parameters achieve their respective thresholds. It is worthwhile to mention that at any point in time, either of the two policy parameters can be considered due to the opposite evolutionary impacts of TFP on these parameters. At any particular value of TFP at time t (between pre- and post-COVID periods), the value of CO₂ emissions is presented as CEt*, and the value of unemployment is denoted as UEt∗. Now, toward the left side of time t, CO₂ emission is rising and unemployment is falling, while toward the right side of time t, CO₂ emission is falling and unemployment is rising beyond the threshold limit. Also, it is observed that both CEt∗ and UEt∗  appear again at time tc  and tu, on the CO₂ emissions–TFP and unemployment–TFP graphs, respectively. Now, if these two scenarios are compared with the ones at time t, then the following conditions can be derived:

Hence, it can be observed that both before and after the occurrence of the turnaround points of the associations, any given value of TFP cannot simultaneously reduce the level of CO₂ emissions and unemployment. Hence, the question of policy trade-off arises. Now, the parallel attainment of SDGs 13 and 8 might be crucial during the post-COVID scenario because of these opposite effects. As none of the associations can attain their optimal values, therefore, a policy optimum needs to be devised to achieve a policy-level solution. This optimum can be obtained at the intersection point of both these associations. In Figure 2, both these associations intersect each other at PI1 and PI2. It signifies that the policy optimum can combine both values and does not require individual optimal values. As innovation is an irreversible process, PI1 cannot be considered the policy optimum. Hence, from a futuristic perspective, the social and environmental impacts of TFP might meet again at PI2, expected to occur during the post-COVID period. Though this point represents neither of the individual optimum values, attainment of this point might be crucial from the perspective of policy-making. The policymakers might be interested in accelerating the achievement of both the turnaround points, followed by delaying the occurrence of PI2. For this reason, both the curves flatten toward the right.

5.1 Implications for theory

This study has made an attempt to extend the famous “Capital-labor substitution” principle proposed by Arrow et al. (1961). The seminal work by Arrow et al. (1961) shows that the economic efficiency in the international trade might be achieved by substituting capital with human labor. It gave an indication that the human laborers will be substituted by the capital following a shape convex to the origin. However, the notion of technological development during that study was limited to the manufacturing sector, which was characteristically labor-intensive. With the advent of technological innovations in the age of Industry 4.0 and growth of the service sector, the notion of technological development has undergone a transformation. Hence, the work of Arrow et al. (1961) might be extended from a convex curve, following the evolutionary nature of technological innovations. Moreover, the “Capital-labor substitution” principle majorly looked into the unidimensional aspect of capital. In the SDG regime, the innovation is expected to play a dual role – in environmental and social dimensions. Therefore, the “Capital-labor substitution” principle required an extension from the socio-ecological tradeoff perspective.

From theoretical point of view, this study has shown how technological innovations can initiate socio-ecological tradeoff. Given these competing objectives, innovation can be viewed as a policy instrument appearing as an environmental boon and a social curse. Ideally, innovation has been perceived as a tool for bringing environmental benefits. The studies by Anthony et al. (2018) and Xie et al. (2021) indicate the environmental benefits of innovation. However, the recent work by Chen et al. (2021) indicates the negative social externalities exerted by innovation. Moreover, the “New Approaches to Economic Challenges” by OECD (2020) talk about reinventing the role of innovation in addressing policy tradeoffs, which are inclined toward managing the social sustainability aspects. From this viewpoint, technological innovation is assumed to have a negative social impact. Now, when the nations are embarking on innovation to achieve the SDGs, the policy dimensions should take care of this aspect. This consideration gives the academic literature a direction about how to visualize innovational duality in policy-making.

In this view, the theoretical framework of EKC hypothesis needs to be mentioned. By far, the literature of energy and environmental economics visualized EKC hypothesis as a tool for measuring environmental impact assessment. However, the capability of EKC hypothesis to encapsulate the evolutionary impact of the policy instruments on the target policy parameter has been largely ignored (Balsalobre-Lorente et al., 2021). The quadratic specification of EKC hypothesis gives a leverage to analyze the nonlinear evolutionary impact over a temporal frame, and hence, this framework can be utilized as a policy forecasting tool. Going beyond the traditional environmental impact assessment, this study has shown the power of EKC hypothesis in analyzing the policy tradeoff. To bring additional insights to the analysis, the cubic specification of the EKC hypothesis can be utilized (Sinha et al., 2019). In that process, capturing the movement of the inflection point appearing between the two turnaround points can uncover further intuitions regarding the possible policy tradeoffs.

5.2 Implications for policymakers

As the innovation influences both CO₂ emissions and unemployment, it is not possible to arrive at the best solution for either of the cases due to the opposite nature of the impacts. Therefore, this policy trade-off requires an optimum solution, which will bring the social and ecological impacts of innovation at a certain equilibrium. Hence, the objective of the policy framework should be focused at achieving this equilibrium during the post-COVID period. Now, this policy framework can be designed following a phase-wise schedule. As the impact of technological innovation will be immediately visible on the CO₂ emissions, therefore the first phase of the policy framework should focus on the ecological impact of innovation. For the innovation to achieve its full potential in CO₂ emissions reduction, the policymakers should adopt appropriate measures to reduce the reluctance of the industrial sector to accept these solutions. Moreover, as continued dependence on the fossil fuel solutions might reduce the potential impact of innovation on the CO₂ emissions, a gradual shift from the nonrenewable to renewable energy sources is also required. A drastic fall in the anthropogenic activities during the COVID-19 outbreak has highlighted the emergence of service sector firms. Now the policymakers might need to bring certain policy interventions for their respective nations towards being recognized as service-oriented economies for retaining the environmental quality. Now, in such a situation, overnight sectoral overhaul might create deterrence to the prevailing economic growth pattern of these countries. Therefore, policymakers might need to use the existing financialization channels in such a way so that the firms can embrace the innovations in a hassle-free manner within a predefined time. For this purpose, the financial institutions can introduce the concept of discriminatory interest rates on loans and advances for availing the innovative solutions, whereas this discrimination might be based on the carbon footprint of the firms. Hence, given a fixed period, the dirtier firms will be compelled to avail the solutions at a higher rate of interest, whereas the cleaner firms will enjoy a lower rate of interest. Thus, the economic system in the countries will encourage the adoption of cleaner energy solutions and innovation. During this finance-driven transition, it should be remembered that the higher bracket of the interest rate should not be higher than the existing average cost of capital of the firms. Otherwise, it might dissuade the firms to embrace the innovation for improving the quality of their production systems. Additionally, they might stop the business operations. Hence, the rate of interest should not be a deterrence in the innovation adoption process.

Once this phase becomes operational, the second phase of the framework should be designed to internalize the negative social externalities. As the rising diffusion of innovation is starting to replace the human labors after the deployment and implementation period, there is a requirement of an early policy intervention. When the firms start implementing the innovative solutions in their existing production processes, there should be an upper threshold of innovation for the firms, beyond which they will not be able to replace human labour with technology. Following equations (3) and (4), the level of TFP can be denoted as per the following [1]:

At the point denoted in equation (11), the level of TFP to be achieved by any firm allows it to tread along the long-run equilibrium growth trajectory. Moving beyond this point might help the firm to gain a short-run economic profit at the cost of employment. Now, the rise in consequential unemployment might create a demand pressure in the economy, leading to the reduction in supply and production. Treading along the long-run equilibrium growth path necessitates the production to continue. For this purpose, certain short-run economic losses need to be incurred. Therefore, the policymakers need to ensure that the policy intervention point should be achieved immediately after the first phase of the policy framework is implemented to keep the economic growth trajectory intact. The accomplishment of the first phase of the policy framework will help these countries in achieving the objectives of SDG 13, whereas the second phase will allow these nations to achieve the objectives of SDG 8. Thus, this policy framework ensures sustainable development in the post-COVID scenario.

While the core policy framework helps these nations to avoid the policy trade-off, the tangential policy framework helps to sustain the core policy framework. After the first two phases of the policy framework being operational, policymakers need to gradually control the population growth rate. If the population growth rate is higher than the growth rate of job opportunities, it creates the unemployment issue. As a result, the core policy framework might not be able to achieve its full potential. Hence, strict population control mechanisms should be devised by policymakers. Moreover, the policymakers should encourage the start-up ventures with these technologies to diffuse the innovations in a more effective manner. However, the capital allocation should be done in such a way so that the firm can generate enough employment, not fully replacing the labor with technology. For institutionalizing these solutions, the educational curriculum should be amended. It can raise the students' awareness of the latest technological developments and innovations across the nations and the social and ecological benefits of innovation. Thus, it will help the nations in accomplishing the objectives of SDG 9 (industry, innovation and infrastructure) and SDG 4 (quality education).

5.3 Policy caveats and assumptions

Discussion of a policy framework is seemingly incomplete without mentioning the caveats and assumptions behind the framework. Understanding these two aspects is necessary, as their non-fulfillment might hinder the policy framework from reaching its full potential (Cheng et al., 2021a, b). First, during bringing discrimination in the interest rate, the slabs should be made close so that many firms can be encapsulated. Second, policymakers need to introduce rehabilitation policies and vocation centers for the labor employed in the traditional fossil fuel-based energy generation sector. Third, the rent-seeking mechanism in the bureaucratic system should be brought to the minimum, as such incidents of corruption hinder the diffusion of innovation within and across the borders.

5.4 Limitations and future projections

Though this study has introduced a critical policy dimension by describing the socio-ecological policy trade-off initiated by innovation, the study might suffer from certain limitations. One of the major limitations of the study is that only the output indicator of innovation is considered, while various other forms of innovation (e.g. social innovation, environmental innovation) have not been incorporated. Though theoretically proven, putting forth a generalized view of innovation might yield different results in the empirical pursuit. Saying about this limitation, this is also needed to clarify that the policy framework introduced in this study can serve as a baseline policy approach for addressing the policy trade-off in any context. Moreover, the framework is flexible to accommodate any additional policy instrument, which might be contextually suitable. This flexibility and generalizability have made this policy framework a contribution to the literature. Future studies in this direction can be carried out by considering various forms of innovation and how those forms can demonstrate socio-ecological trade-offs in various contexts.