Factors related to information and communication technologies adoption in small businesses in Colombia

This study is an analysis of how a set of factors influenced the decision to adopt information and communication technologies (ICTs) in small businesses in Colombia; for this purpose, a logistic modeling is applied, wherein the explained variable represents the companies’ decision to adopt ICTs. The main explanatory elements include demographic aspects of the owner, the company characteristics, and the legal practice of the company and its social capital. Data from the Microbusiness Survey 2019 conducted by the National Administrative Department of Statistics of Colombia were employed. The results show that traits such as owner age and gender affect adoption. Furthermore, variables such as the age and size of the company, the amount of computer equipment, business formalization practices, and social capital influence the use of these technologies. Therefore, it follows that the individual features of business owners are also determinants of technology adoption, and their inclusion becomes essential to supplement the analysis of these issues. In addition, the study supports the evidence that the acceptance of these technologies is positively related to the business size and infrastructure. By exploring, for example, the business formalization factor, a positive effect is obtained; consequently, the implementation of legal practices is linked to the ICT use. Thus, some of the areas for policy implementation would be, on one hand, the endowment of computer equipment, since it is a critical aspect in these small businesses, for which a subsidy policy for the acquisition of devices would be useful. In addition, the promotion of business formalization practices could lead to an improvement in the firm’s culture of technology use.


Introduction
In today's techno-economic world, information and communications technologies (ICTs) are generally regarded as a crucial tool to improve the competitiveness of the economy at micro and macroeconomic levels.The consensus in the extant literature is that ICTs have a significant influence on company productivity, but this influence will depend on the extent to which they are disseminated and widely used (Brynjolfsson et al., 2021;Pieri et al., 2018).For this reason, understanding the determinants of providing a framework of evidence on what aspects are related to a greater acceptance of ICT use, and as a consequence, it suggests policy lines in the direction indicated by the results.
The remainder of the paper is organized in the following manner."Background and Theoretical Framework" section presents the main background and theoretical framework."Methodology" section presents the methodology employed, describing the database and the empirical model used."Results and Discussion" section states the results and discussion."Conclusions" section mentions the conclusions.

Background and theoretical framework
The adoption of technologies in the framework of technology, economy, and society studies is linked with the concept of diffusion as a fundamental process to ensure their interaction with economic growth.For this to occur, it is necessary for companies to be able to absorb and successfully apply new technologies.Thus, it is essential to identify the determinants of this adoption and diffusion both from the research and technological development policy options viewpoints.
Within the approaches of technological adoption, it is established that this is more of a social process that includes technical-social entities that shape the adoption behavior involving multilayered situations or contexts including structural arrangements, culture, division of labor, specialization, standardization, operating procedures, communication patterns, ideologies, government regulations, competitive forces, supplier strategies and knowledge about technology, and socioeconomic conditions, (Chen & Chang, 2013;Jacobsson & Linderoth, 2010;Orlikowski, 1992;Venkatesh et al., 2003).In relation to the role of the business environment, Jacobsson and Linderoth (2010) stated that it can affect both the adoption and implementation of new ICT applications and the existing power relations in the organizations.
In this section, we intend to move toward a theoretical-empirical framework that allows us to analyze some of the factors that affect the decision to adopt ICTs, especially in small companies, and given the data limitations that do not allow us to have a wider range of elements to consider in the empirical exercise.For example, a key factor not incorporated here is the role of business strategies, Giotopoulos et al. (2022), and Molinillo and Japutra (2017) argue that companies that prioritize strategies that are strongly linked to marketing or customer needs are more likely to drive ICT adoption.
In this context, the aspects to include and whose influence is to be tested in the adoption process are demographic characteristics of the owner or manager (gender and age), characteristics of the company (age, size, and computer equipment), legal practices of the company (formality), social capital (associativity and cooperative affiliation), and the diffusion or epidemic effect of ICTs (Fig. 1).The analysis of the variables and their influence is explained below.

Demographic characteristics of the owner
Some studies have concluded that the demographic characteristics of the owner or manager of a company affect the adoption of technologies, either directly or indirectly (Abdullah et al., 2012;Chuang et al., 2009;Naicker & Van Der Merwe, 2018;Peltier et al., 2012;Viollaz, 2018).In the case of age, the willingness to adopt would be inversely correlated, i.e., young owners/managers show more openness toward the application of new knowledge and technologies, and in relation to gender, some studies suggest that men would have a greater interest in analyzing and adopting ICTs, while women, given their reluctance to be trained in computer skills, seem to be less willing to adopt these technologies (; Ilie et al., 2005;Harrison & Rainer, 1992).

Age
Although there is no consensus about its impact on adoption, it should be highlighted that most of the studies are in favor of a positive effect on this decision.Thus, greater productive experience reflected in company durability would have an impact on embracing ICTs, because companies would have a better vision of the need and advantages offered by new technologies, as opposed to the uncertainty or disinterest that younger companies may show (Khalifa, 2016;Meyer, 2011;Viollaz, 2018;Youssef et al., 2010;Dunne, 1994).As an exceptional case, this variable does not demonstrate any statistical significance in Buenrostro and Hernandez (2019).Giuri et al. (2005) confirmed that the level of ICT investments, the total amount, and the amount per employee increase with the size of the company; moreover, large companies have a higher proportion of telecommunications investments.The four following reasons explain this positive relationship: first, large companies have a greater capacity to own the benefits of adoption.Second, large and more profitable companies will also have easier access to the financial resources needed to purchase and install new technology.Third, large companies have the necessary skills and other resources to introduce and use new technologies.Fourth, as suggested by Khalifa (2016), large companies can quickly capture economies of scale of production through the learning curve and can spread the other fixed costs associated with adoption in number of units.For example, studies such as those of Loukis et al. (2017); Viollaz (2018); Youssef et al. (2011);and Galliano and Roux (2008); Lucchetti and Sterlacchini (2004), reveal the positive effects of the size of the company in the adoption of technologies.Finally, we can to mention Hollenstiein (2004) and Meyer (2011) as cases in which there is no evidence of this positive effect.

Computer equipment
This factor refers to the amount of ICT-related equipment that the company has for its production, administration, and business management activities.It is related to the number of devices, but it is also very intricately linked to the sophistication of the ICT infrastructure, because the effect on adoption is expected to be positive.In this regard, Loukis et al. (2017), based on the technology, organization and environment theory of Baker (2011), conduct a study, where they observed a strong relationship with positive effects between technology adoption and sophistication and the adoption of cloud computing, the latter being an indicator of ICT adoption for the glass, ceramics, and cement sectors in the European Union.

Legal practices
Formalization is the process conducted by companies to join the formal economy.According to International Labour Organization OIT (2017), the main characteristics that distinguish business formalization are being registered and licensed by national, provincial, and municipal authorities, thus having access to social security for the owner of the company and its employees, in compliance with the current legal system, including tax and labor legislation.
On the contrary, lack of formality (informality) in companies is related to the development of production and/or sales while avoiding state regulations that are deemed as a source of business inefficiency, because the payment of taxes is considered to be unnecessary to achieve business goals.Moreover, the relation that there is between companies being sized and the presence of informality is widely known (Gómez & Moran, 2012), as small companies are the ones that find greater difficulties in complying with the legal requirements to access formality.
Informal sector companies usually suffer from having low levels of technology, a phenomenon that has contributed to their poor development, and the causal relationship at first glance seems quite clear, because the consequences are low productivity and little progress in productive and organizational terms.As the two defining characteristics of ICTs are their convergence and their speed, there is a wide range of possibilities for the collection, manipulation, transmission, storage, and presentation of information; through these possibilities a completely new way of conducting business is created.Consequently, small informal enterprises are more vulnerable to the new economic context governed by these technologies.In this regard, a study by Opiyo and K' Akumu (2006) on the assessment of the role of ICTs in the development of the informal sector in the Kariokor market cluster in Nairobi concluded that there is a need to prevent further marginalization of the informal sector through the use of computer and digital technologies.Along these lines, the relationship between formality and ICT adoption could be established based on business practices and culture.If companies have better legal habits, it is possible that they also develop appropriate management policies, including the monitoring and adoption of technological trends in the sector.

Social capital
Social capital is understood as the element of human capital that allows members of a given society to trust each other and cooperate in the formation of new groups and associations, (Cohen & Prusak, 2001;Dasgupta, 2000;Rupasingha et al., 2006).The ability to form organizations depends on concepts, such as property rights, contracts, and a system of commercial law.However, it also depends on a prior sense of moral community, that is, an unwritten set of rules or ethical norms that constitute the basis for social trust.Trust can dramatically reduce transaction costs, bargaining costs, and enforcement and makes possible certain efficient forms of economic organization that would otherwise be blocked by extensive rules, contracts, litigation, and bureaucracy.Moral communities, as lived and experienced by their members, tend to be the product not of rational choice in the economists' sense of the term, but of nonrational habit.Consequently, one would expect that through the bonds of trust and interaction between companies, it would be possible to transmit information about the benefits of and need to adopt certain technologies in the productive sector, so that companies with greater social capital will have access to better information that would lead them to make the decision to adopt ICTs.In this sense, the work of Lee et al. (2013) showed the influence of social capital in the process of tourism technology adoption, where the trust factor played an important role in the dialogue and transmission of information between managers.

Diffusion or epidemic effect
The studies of Haller and Siedschlag (2011) and Windrum and Berranger (2003) find that ICTs have greater diffusion in the most knowledge-intensive activities.Therefore, it is essential to consider the effect of the diffusion rate of the industry or economic conglomerate to which the firm belongs.In this regard, Rogers (2003) suggests the importance of the overall diffusion of technology in the individual firm's adoption decision; the higher the diffusion of a technology or innovation is, the higher the probability of its adoption by the marginal firm is, then would indicate that the technology is being widely accepted by social agents.

Data
The empirical basis of this study is the cross-sectional data of the Microbusiness Survey 2019, conducted by the National Department of Statistics, which is the official statistics office of Colombia.This survey collects information from the smallest size companies in Colombia, with less than 10 employees; thus, it is a tool focused on capturing data from small businesses, which due to their large number and dynamics are not included in other surveys on the industry or trade conducted by the Colombian government.
Among the main issues addressed in the survey, there are aspects, such as company characteristics, owner characteristics, business formality, entrepreneurship, use of ICT, social capital, and employed personnel.For the year 2019, the survey included a sample of almost 87,000 small companies from the following sectors: agriculture, manufacturing, commerce, and services.The data to be used are at the company level, so it is feasible to approach a microeconometric modeling.

Model and variables
To establish the factors affecting ICT adoption a logistic regression model was planned, where the dependent variable takes the value of 1 if the company has adopted ICTs, and 0 if it has not.As for the adoption variable, we chose to assume as adopters those companies that reported having access to and using the Internet, since this would imply that the business has incorporated the use of this technology in its routine activities to a greater or lesser extent.While the independent variables refer to: the characteristics of the owner (age, gender), the characteristics of the business (age range, business formality, number of employees), equipment, social capital of the business (affiliated to associations, affiliated to a cooperative), economic environment (productive sector to which it belongs, the business is located in one of the main cities of the country) and the adoption rate of the economic sector to which it belongs.Table 1 shows the definition of each of the variables to be used and suggests the expected sign of the parameter according to the theoretical framework and literature.
The choice of the model has been based on the availability of information in the data used, as well as on the review of the literature, where it can be observed that an adequate way to explain the decisions of technological adoption is through binary choice models, such as logistic regression, since it is assumed that the firms will make a choice between two possible alternatives (adopt or not adopt the technology), a decision that depends on the observable characteristics, (Khalifa, 2016;Li et al., 2019).Therefore, the specification is where The term u i represents random error.Table 2 shows the proportions and mean values of the variables that make up the logistic model.It is important to note that the sample includes 86,969 small companies distributed in the main geographic areas and economic sectors of Colombia.For example, 27% of the companies are ICT adopters, a figure that indicates that there is still a larger proportion of businesses that do not use these technologies and provide a basis for (1) (2)

Results and discussion
The estimation of Eq. ( 1) is presented in Table 3, which was calculated with robust standard errors in the coefficients to correct the presence of heteroscedasticity.Column 2 reports the coefficients and column 4 the average marginal effects associated with each explanatory factor.The results indicate that all variables were statistically significant, except for the Medellin dummy one.The majority of impacts of the variables are consistent with the expected signs that were proposed initially.First, gender has a negative sign, indicating that being a male owner reduces the probability of adoption, and in the fourth column the average marginal effect indicates that being male decreases the probability of adoption by 4.9%.The age of the owner also has a negative sign, thus, the older they are, the lower the probability of adoption, since for each additional year the probability decreases by 0.7%.Regarding the previous variables, in the research by Viollaz (2018), it is found that the age of the company administrator does not show statistical significance, and if he is male, the probability increases.This last finding is different from the result obtained here.Therefore, the result regarding gender is contrary to that suggested by previous research (Ilie et al., 2005), where women would be less interested in adoption processes.While regarding the age of the owner, the finding is consistent with the idea that an older owner/manager is less likely to adopt (Abdullah et al., 2012;Chuang et al., 2009).
Regarding the dichotomous variable of company age, the established ranges have a positive relationship with adoption.Note that the reference category is less than 1 year, thus younger companies are less prone to adopting.The positive impact of company age is consistent with the works of Meyer (2011) and Viollaz (2018), while Buenrostro and Hernandez (2019) did not show statistical significance.The number of employees has a positive effect, thus, the greater the number of employees working in the company, the greater the willingness to embrace ICTs.This result is aligned with research by Lucchetti and Sterlacchini (2004), Galliano and Roux (2008), Battisti and Stoneman (2005), Khalifa (2016), Loukis et al. (2017) and Viollaz (2018), where it is shown that the larger the company, usually measured through the number of employees, the greater the willingness there is to adopt computer and communication technologies, and this is contrary to the work of Hollenstiein (2004) and Meyer (2011).Overall, the evidence leans toward the theory that adoption increases with company size.
The amount of computer equipment owned by the establishment also shows a positive effect, and for each additional piece of equipment, the probability of adoption increases by 23.5%, thus evidencing the importance of equipment provision for the use of such technology.Therefore, this is consistent with the study of Khalifa (2016), while according to Loukis et al. (2017) the positive effect is associated with the sophistication of ICT infrastructure, but the result of Buenrostro and Hernandez (2019) is ambiguous.
Business formality also makes it more likely that the decision to adopt the technology is taken, indicating that legally registered businesses also tend to use these new technologies more.In this case, the average marginal effect yields a value of 9.2%.Moreover, social capital variables such as being associated and affiliated to a cooperative have a positive impact, with marginal effects of 6.4% and 12.2%, respectively.Thus, associativity, in addition to strengthening trust between companies, also promotes trust in the practices developed by them, and it allegedly improves acceptance in the use of ICTs.
In addition, the company's adoption of technology is positively affected by the adoption rate of the economic sector of the business, and it may be influenced by the aggregate behavior of the sector, thus spillover effects could be seen in the diffusion of these technologies.This result is consistent with the research of Hollenstiein ( 2004), Galliano and Roux (2008) and Khalifa (2016), thus reinforcing the hypothesis of spillovers in the use of these technologies.
Conversely, in the dichotomous variables of the three established economic sectors positive signs are obtained, here the reference category is agriculture.Thus, belonging to the three increases adoption with respect to agricultural companies.
Finally, among the dichotomous variables for the main cities, Cali shows a positive sign, while Bogota shows a negative sign.Thus, small companies in Cali tend more to adopt the technology and in Bogota less so.

Conclusions
The study shows that demographic aspects associated with the characteristics of the company's owner, which have been less addressed in the extant literature on ICT adoption, are statistically significant.For example, age has a negative impact, as does gender (male).In this case, as they are small companies depending on few people, the characteristics of the owner or manager play an important role in business policies, and hence, these factors should be considered relevant in adoption studies.
Another research finding supports the evidence that the age or longevity of the business positively affects adoption.Therefore, as companies gain maturity or experience, they see technology as a necessary tool for their sustainability.In addition, company characteristics such as size, approached through the number of workers, and the amount of ICT equipment have a positive impact on adoption.Additionally, business formality practices are correlated with a greater probability of using ICT, since operating with formal legal practices would be associated with better administrative habits, which recognize the need to implement new technologies.
In addition, factors associated with the company's social capital, such as membership in productive associations and cooperatives, increase the probability of technology adoption, and the bonds of trust and the sharing of information between companies generated in these networks can help them adopt the new technologies.
Furthermore, it would be desirable to conduct research on the use of ICTs to address other problems in small enterprises in developing countries.For example, exploring the effects of the use of these technologies on productivity, profitability, and complementarity with other aspects of business, which would highlight the impact they have on these economies.

Policy implications
According to the results, some of the potential policy alternatives are, first, the endowment of computer equipment, because the poor availability of computer infrastructure in small businesses in Colombia, is undeniable.Therefore, a subsidy policy to acquire devices would be convenient.Second, government and private-sector entities, such as Chambers of Commerce, must design strategies to promote the practice of business formalization in this sector, because there is evidence that it leads to the improvement in the firm's culture of technology use as well.Furthermore, as Table 2 shows, barely 13% of enterprises are formal; consequently, increasing the percentage of businesses operating within legal frameworks becomes a goal.

ICTs
Information and communication technologies

Fig. 1
Fig. 1 Determinants of ICTs adoption.Source: Prepared by the authors β 3 owners ′ age i + α j j=1 business ′ age ji + β 4 formality i + β 5 employees i + β 6 amount equipment i + β 7 associated i + β 7 cooperative affiliate i + γ k 10% of the companies analyzed are located in the three main cities of Colombia, and the remaining 90% are in the rest of the country.At the bottom of Table2, are the averages of the numeric variables.The figures show that, on average, a business owner is 46.5 years, which represents a mature age level.The number of employees per establishment is 1.4, which can be considered relatively low.The amount of equipment for ICT use per company is 0.3, revealing little adoption of these technologies in this business segment in Colombia.

Table 2
Average proportions and values of the variables

Table 3
Results of the logistic model of adoption