Technological capabilities refer to a firm's ability to develop new products and services by aligning its strategy with innovative processes (Wang, 2007). Such capabilities involve knowledge and skills in acquiring, using, absorbing, adapting, improving, and generating new technologies (Bell & Pavitt, 1995Malhotra, Mathur, Diddi & Sagar, 2021;). These capabilities have enabled the development of new products and technologies, improving manufacturing processes and quality control skills, and predicting technological changes in the industry (DeSarbo, 2005).

Similarly, as a system, digital capabilities are provided through another supplier-user interaction that creates value, also defined as a digital output or service (Srivastava & Shainesh, 2015). Kohli and Grover (2008) defines it as the internal capability to provide customer information at the required instant. According to Lyytinen, Yoo and Boland (2016), digital capabilities are digital systems that generate new outcomes and structures without uncoordinated participation of third-party actors and without deliberate planning by the system creator. Consistent with Yoo, Boland, Lyytinen and Majchrzak (2012), multiple products or subsystems can be designed and controlled using the same digital tools.

Technological capabilities have enabled organizations to create opportunities to support their competitive advantage. In this regard, understanding the dynamics that influence the adoption of digital technologies becomes critical to their success El-Haddadeh (2020). examines the specific aspects of digitization that affect technology adoption in organizations Karimi and Walter (2015). analyze the impact of dynamic capabilities on digital disruption in firm performance. Digital disruption reduces intermediation costs (Sutherland, 2018) and incorporates technology more efficiently (Karimi & Walter, 2015).

In this sense, recent empirical studies have left the relationship between digital capabilities and firm performance ambiguous Torres, Sidorova and Jones (2018). highlight the role of business intelligence and analytics (BI&A) and its positive connection with firm performance from the perspective of dynamic capabilities. However, Usai et al. (2021) and Chen, Lin, Chen, Chao and Pandia (2021) propose that companies reshape their capabilities to meet new market demands by adopting digital transformation as their primary organizational strategy. This environment can also be presented to IT capabilities (Grewal, Comer and Mehta, 2001), artificial intelligence and machine learning (Gordini & Veglio, 2017), use of virtual reality (Boyd & Koles, 2019), video conferencing (Hardwick & Anderson, 2019), and the internet (Avlonitis & Karayanni, 2000), among others, that act as reconfiguring capabilities and resources within an organization. Given the reconfiguring role of digital capabilities in other capabilities, we proposed that digital capabilities, as third-order capabilities, do not positively affect firm performance; instead, this relationship can be harmful.

Thus far, there is evidence that digital dynamic capabilities, which complement other capabilities, have a positive relationship with firm performance Soluk, Miroshnychenko, Kammerlander and De Massis (2021). studied the role of dynamic capabilities as mediators in the relationship between family influence and digital business model innovation (BMI) and the moderating role of environmental dynamism. Similarly, Khin and Ho (2019) examined the effect of digital orientation and capability on digital innovation and their mediating impact on the link between firm performance, digital direction, and digital capability. However, the present study includes new variables such as management and digital capabilities. Additionally, the interaction of HDI (moderating variable) with digital capabilities, technological capabilities, and firm performance has emerged in response to the “new normal.”

The fundamentals of ensuring a successful adoption process depend on how organizations can identify the need to adopt such technological innovations (Kim, 2015). Hence, there is a need to adopt digital tools to improve the effectiveness of organizational functions and processes (Osmonbekov, Bello & Gilliland, 2002). Digital capabilities allow us to take advantage of a large amount of information from the environment and incorporate the technology more efficiently according to an organization's value proposition (Gobble, 2018).

In this “new normal,” Almeida, Santos and Monteiro (2020); propose indications of the impact of digital capabilities in three business areas in the new normality: labor and social relations, marketing capabilities, and technological capabilities. Based on the above, digital technology combines technological capabilities to meet an organization's objectives.

Digital tools have allowed companies to permanently capture significant volumes of information regarding consumers in recent years. In this sense, the development of digital capabilities acquires relevance in organizations, allowing the use of information to reduce the cost structure and redesign their processes. Consequently, technological capabilities are activated by digital ones, improving the quality of products and services provided and, therefore, improving firm performance (Ciampi, Demi, Magrini, Marzi & Papa, 2021).

This study considered that digital capabilities enable the identification of signals from the environment and as the main complement to other capabilities. Likewise, third-order capabilities allow for the activation of resources and inferior capabilities. Thus, the following hypothesis is proposed:

H1: Technological capabilities have a mediating effect between digital capabilities and firm performance.

Collaborative economies are emerging and growing through the possibility of developing or transforming new activities in emerging economies. Collaborative economies help to reconfigure capabilities and resources through digital technologies (Malik, 2020). Similarly, developing new capabilities in companies can lead to disruptive opportunities Fainshmidt, Pezeshkan, Lance Frazier, Nair and Markowski (2016). argue that new capabilities contribute more significantly to firm performance in emerging economies than in developed ones. Institutional factors such as informality and institutional gaps promote collaborative economies in emerging markets (Boateng, Kosiba & Okoe, 2019Heredia, Geldes, Kunc & Flores, 2019;).

Therefore, collaborative economies have more significant advantages in emerging economies because of their lower levels of investment and more straightforward implementation (Dokko, 2015). Thus, in economies with a low level of development, the use of collaborative economies significantly reduces intermediation costs (Dokko, 2015).

Thus, it can be said that emerging economies have a low HDI. The initial premise of this index was that national development should include both life expectancy and literacy. Amartya Sen is the pioneer of this approach, and it is a widely known concept of human welfare or development. Emerging economies with a low level of HDI present high levels of creativity, and collaborative economies are growing in emerging economies (Vega-Muñoz, Bustamante-Pavez and Salazar-Sepúlveda, 2019). The development of a collaborative economy in emerging economies requires less complexity in the design of these tools than in developed economies. HDI represents the level of public health infrastructure (Ross & Wu, 1996), inequality, and poverty (Leung, Sharma, Adithipyangkul & Hosie, 2020).

During the pandemic, the term “gig economy” became popular due to the growth of digital platforms and apps. In this sense, the rapid spread of digital technology has generated new economic activities. A study argues that the pandemic positively impacted the “gig economy” (Umar, Xu & Mirza, 2021). In addition, communication technologies, digital platforms, and apps have allowed people with a computer, cell phone, and internet connection to perform jobs from anywhere in the world (Anwar & Graham, 2021). Therefore, intensification of the gig economy in emerging countries with a low HDI generates new employment opportunities (PNUD, 2020). An essential aspect of the “gig economy” is that it offers non-traditional forms of work because there is no employer-employee relationship, achieving greater flexibility, as workers decide the hours and days they work.

The rapid growth of the “gig economy” in emerging economies (low HDI) is because there is a greater supply of informal workers in these economies as well as a greater adoption of temporary work by mobile applications driven businesses. In this sense, companies’ digital capabilities increase, reducing business costs and increasing firm performance. Thus, we propose the following hypothesis:

H2: In countries with low HDI, the impact of digital skills on technological capabilities is more significant than in countries with high HDI.

The present study considered four variables through internal factors (resources and capabilities) and external factors (moderating HDI) that are relevant to firm performance in this “new normal.”

The independent variable of the model was management capability. The mediating variables of the model were digital and technological capabilities. Finally, the dependent variable was firm performance. We emphasize that all the variables are constructs in which a series of indicators have been formed Table 1. shows descriptions of these variables.

We used the “Enterprise Survey – Covid19: Impact on firms.” The World Bank designed this study to measure the effect of the virus on the countries considered in the sample. The measures used for each variable in the study were dichotomous. The constructs used were management capabilities, technological capacity, digital capabilities, and firm performance. The liquidity variable represents management capability. The measure of technological capabilities is the variation in operating hours and supplies—digital capabilities measured by online activity, delivery, and the adoption of remote work. Finally, firm performance refers to two items: change in sales and demand for goods and services (see Table 1).

All constructs were measured as mode A composites (Henseler & Schuberth, 2020). We analyzed the standard method to assess the quality of the information, which could improve the relationships between variables when collected from the same source. The analysis applying Harman's single-factor test (Podsakoff & Organ, 1986) did not reveal that the variables were grouped into a single factor, thus presenting no problems (Twigg, Kutzer, Jacob & Seaman, 2019).

We included the following control measures to eliminate the possible endogeneity of the omitted variable bias (Hamilton & Nickerson, 2003), which has been shown in the literature to impact firm performance. Size is measured using the number of employees in a firm (Clauss et al., 2021Medase & Abdul-Basit, 2020; Moretti & Biancardi, 2020;). We also included industry-type dummy variables to represent the selected subsectors (manufacturing, retail, and services) (Clauss et al., 2021Höflinger, Nagel & Sandner, 2018;). To detect and overcome the possible existence of unobserved heterogeneity bias due to unknown factors, we followed the proposal of Motiwalla, Albashrawi and Kartal (2019) and segmented the sample into different groups according to countries with high or low HDI.

We tested our research model using partial least squares (PLS) and the SmartPLS package (v.3.3.3) (Ringle & Sarstedt, 2016). We used PLS for the following reasons.

(1) The research model is complex, depending on the type of hypothesized relationships (direct, mediation, and moderation); (2) It has the advantage of not imposing distributional assumptions for the indicators, and (3) The constructs included are composite (Chin, 2010Sarstedt, Henseler & Ringle, 2011;). This model is also widely used to test theories using the dynamic capabilities approach (Ali, Javed & Danish, 2021Ebrahimi, Hamza, Gorgenyi-Hegyes, Zarea & Fekete-Farkas, 2021; Frempong, Mu, Adu-Yeboah, Hossin & Adu-Gyamfi, 2021;).

The application of the PLS technique involves several steps, including model fitting, applying a bootstrapping process (10,000 subsamples), following bootstrap-based exact fit tests for the estimated model (Henseler, Ringle and Sarstedt, 2016). Further, we evaluated the measurement model by analyzing the fit of the saturated model (Müller, Schuberth and Henseler, 2018). Finally, the algebraic sign, magnitude, and statistical significance of the path coefficients were assessed and coefficients of determination (R2) were evaluated (Ali, Rasoolimanesh, Sarstedt, Ringle & Ryu, 2018).

We tested the moderating hypothesis, H2, using multi-group analysis (Henseler & Fassott, 2010). We divided the sample into two groups: countries with a high or low HDI. Before comparing the path estimates between groups, it was necessary to ensure measurement invariance of the composites. Then, it was possible to ensure that the effect of HDI as a moderating variable in the model was due to the path coefficients of the structural model and not because of the parameters of the external model. A three-step procedure for analyzing the measurement invariance of composite models (MICOM) was used (Henseler, Ringle and Sarstedt, 2016).

The indicator loadings of each construct were generally above 0.707, except for the online activity indicator, whose value was 0.664, which was slightly below the threshold. We retained to improve the explanation of the construct (Henseler & Schuberth, 2020). The composite reliability of the constructs was above 0.7, and AVE was above 0.5. Cronbach's alpha and rhoA values for the technological and digital capabilities constructs were slightly below the cutoff. However, we kept these indicators considering that a value above 0.5 may be sufficient (Taherdoost, 2016). We evaluated discriminant validity according to the Fornell-Larcker criterion (Fornell & Larcker, 1981) and heterotrait-monotrait ratio (HTMT) value (Henseler, Ringle and Sarstedt, 2016). These values are listed in Table 2.

Following Hair, Sarstedt, Ringle and Gudergan (2017), to assess the statistical significance of the path coefficients, we used bootstrapping (10,000 resamples) to generate t-statistics, along with confidence intervals. The R2 values for firm performance and technological capability were significant Fig. 1. shows the results of this model. In total, only one of the direct effects was non-significant, whereas the other five were significant.

Finally, we followed the proposed measures to assess the goodness-of-fit of the estimated model (Dijkstra & Henseler, 2015). The standardized root mean square residual (SRMR) of the model was less than 0.10, as proposed in other studies (Ringle, Sarstedt & Straub, 2012Williams, Vandenberg & Edwards, 2009;) Table 3. shows that the variances were not significant because the 99 percent bootstrap quantiles of the value of the three measures (SRMR, the unweighted least squares discrepancy dULS, and the geodesic discrepancy dG) were higher than the original values (Henseler, Ringle and Sarstedt, 2016), making the model explanatory.

To test the mediation hypothesis, H1, we applied the analytical approach described by Nitzl, Roldan and Cepeda (2016). We specified the indirect effect of digital capabilities on firm performance through technological capabilities. As shown in Table 4, digital capabilities do not directly affect firm performance. As (Usai et al., 2021) mentioned, we identified complete mediation by introducing technological capabilities, supporting the idea of Wang (2007) that technological capabilities complement digital capabilities. In addition, the results did not reveal heterogeneity bias for any of the controlled variables. Thus, Hypothesis 1 is accepted.

To test the moderating hypothesis, H2, concerning multi-group analysis, we used a measurement model invariance of composite models (MICOM) (Henseler, Ringle and Sarstedt, 2016). This approach involved three steps: (i) configuration invariance, (ii) compositional invariance, and (iii) full-measurement model invariance (Henseler, Ringle and Sarstedt, 2016; Sarstedt, Henseler and Ringle, 2011). We used the MICOM procedure by running a two-tailed permutation test for the moderating variable (HDI) at a 5% significance level and 10,000 permutations. The results of the MICOM procedure are shown in Table 5. The results of the permutation-based multigroup analysis are shown in Table 6, indicating the hypothesis tested for each group as H2a, H2b, H2c, H2d, H2e, and H2f. According to the HDI, three relationships show differences between countries: the effect of technological capabilities on firm performance, the impact of digital capabilities on technological capabilities, and the effect of liquidity on firm performance.

Finally, technological capabilities have a mediating effect between digital capabilities and firm performance at 100% because this study shows that the relationship between digital capabilities and firm performance is not significant. Similarly, through an analysis of impacts comparing countries with low HDI with countries with high HDI, we can affirm that in countries where the HDI is low, the impact of digital capabilities is more significant on firm performance than in countries where the HDI is high. The results do not indicate an unobserved level of heterogeneous bias because the model results did not change when analyzing groups with control variables. Thus, Hypothesis 2 is accepted.

The results show that in countries with a low level of development, collaborative economies have a greater impact because of low costs in these economies. Therefore, we recommend taking advantage of the returns of collaborative economies in developing countries and the increasing demand for activities related to e-commerce because the effect on them is disruptive. Likewise, the context of informality (Boateng, Kosiba and Okoe, 2019;Pérez, Yang, Bai, Flores & Heredia, 2019;) is ideal for developing collaborative economies, expanding products and services, improving market access, and income generation.

Digital transformation can be a critical tool for companies to accelerate technological innovation (Troise, Corvello, Ghobadian & O'Regan, 2022). Our study emphasizes that digital capabilities can generate technological innovation faster than traditional capabilities (long-term resource investment and funding constraints). Thus, companies must digitize as quickly as possible to become more efficient. Additionally, the government should provide facilities for companies to develop technologies more quickly (Heredia, Rubiños, Vega, Heredia & Flores, 2022).

The analysis conducted in our research collected cross-sectional information at a specific point within the context of the “new normal.” In the future, we can analyze firm dynamics and the influence of technology collaboration networks, and how they are affected by macroeconomic cycles, industry life cycles, and firm age (Ramírez-Alesón & Fernández-Olmos, 2021). Also, we propose to analyze the interaction of digital capabilities with other types of capabilities and resources in organizations (such as labor and social relations, marketing capabilities, and technological capabilities) (Almeida, Santos and Monteiro, 2020). Future research should focus on a longitudinal study or conduct analyses using a non-recursive structural model, which shows that a firm's performance influences its resources and capabilities. We propose new moderating variables, such as the globalization index (KOF) and the global competitiveness index (GCI), as suggested by Skare and Soriano (2021). Likewise, another line of future research is on the dynamic perspective of the model, which would allow us to contrast our results and make significant contributions in the light of the appearance of new variants of the COVID-19 virus.

Finally, future research should focus on replicating this study in developed and vaccine-producing countries, given that our study only considered a segment of vaccine-receiving countries to monitor the effects of the pandemic (Dinleyici, Borrow, Safadi, van Damme & Munoz, 2021).