Digitalization and the investment in the technical skills of Finance
professionals on the effectiveness of the Finance function
Name: Lieke Dirckx Student number: 13348582 Date: June 20, 2022
Word count: 12.784
Supervisor: dhr. Prof. Dr. F.H.M. Verbeeten MBA MSc Accountancy & Control, variant Control EBEC code: EC 20211026081055
Faculty of Economics and Business, University of Amsterdam
Statement of Originality
This document is written by student Lieke Dirckx who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
Along with the rise of big data, a new role for controllers has emerged whereby the controller will be in charge of data preparation, analytics, and applications. Prior studies have shown that this results in a skills gap for controllers’ expertise in technology tools. Drawing on the absorptive capacity theory of the Finance function and recent literature, this study first investigates the impact of digitalization in Finance on the effectiveness of the Finance function.
Second, this study assesses the impact of investment in the technical skills of Finance professionals on the effectiveness of the Finance function. Then this study looks at whether the investment in the technical skills of Finance professionals strengthens the relationship between digitalization and the effectiveness of the Finance function. The results of a survey of 182 respondents demonstrate that the direct effects of digitalization in Finance and investing in technical skills are significantly related to the effectiveness of the Finance function. Investing in technical skills, however, does not strengthen the relationship between digitalization in Finance and the effectiveness of the Finance function.
Contents
1 Introduction ... 6
1.1 Background ... 6
1.2 Research question ... 8
1.3 Reading guide ... 9
2 Literature review ... 10
2.1 Absorptive capacity ... 10
2.2 Effectiveness of the Finance function ... 11
2.3 Digitalization in Finance ... 12
2.4 Technical skills ... 15
2.5 Hypothesis development ... 16
3 Research methodology ... 20
3.1 Research design ... 20
3.2 Sample characteristics ... 20
3.3 Variable measurement... 21
3.3.1 Dependent variable ... 21
3.3.2 Independent variable ... 22
3.3.3 Control variables ... 24
3.3.4 Summary variables ... 26
4 Results ... 27
4.1 Descriptives ... 27
4.2 Correlations ... 28
4.3 Regression analysis ... 30
4.3.1 Hypothesis I ... 31
4.3.2 Hypothesis II ... 32
4.3.3 Hypothesis III ... 33
4.4 Additional analysis ... 35
4.4.1 Digitalization in Finance ... 35
4.4.2 Firm size ... 37
4.4.3 Analysis per sector ... 38
4.5 Robustness check ... 39
5 Summary and conclusions ... 41
5.1 Summary ... 41
5.2 Conclusions ... 42
5.3 Limitations and future research ... 44
References ... 46
Appendices ... 51
Appendix A – Factor analysis and Cronbach Alpha analysis of the effectiveness of the Finance function ... 51
Appendix B – Factor analysis and Cronbach Alpha analysis of digitalization in Finance ... 52
Appendix C – Factor analysis and Cronbach Alpha analysis of investment in technical skills of Finance professionals ... 53
Appendix D – Spearman correlation test ... 54
Appendix E – VIF values multicollinearity ... 55
Appendix F – Factor analysis split-up of DIG_FIN ... 56
Appendix G – Robustness check ... 57
1 Introduction
1.1 Background
The term ‘Big Data’ may be traced back to the mid-1990s when John Mashey, former Silicon Graphics (SGI) employee, acknowledged big data as a concept as well as a phenomenon (Diebold, 2012). From this time onwards the big data trend continued, and the discipline gained traction. Big data is described as information that is too massive and complicated to acquire, process, and analyze using present computing facilities. It is often described as the utilization of big volumes, high velocity, a variety of data sources, veracity, and value (Gudivada et al., 2015). In 2010, Eric Schmidt, Google’s CEO at the time, said at a conference that the quantity of data generated in two days is comparable to the total quantity of data generated from the origin of civilization to the year 2003 (Siegler, 2010). Since then, the amount of data produced has increased even more. This expansion is known as ‘big data’, and it has cleared the path for better financial decision-making. Big data provides new opportunities which can be used by controllers to make more appropriate and timelier decisions about their business (Brands &
Holtzblatt, 2015).
Gathering big data is relatively simple, but processing and extracting valuable information from vast amounts of data is not seen as that simple. Along with the rise of big data, a new role for controllers has emerged whereby the controller will be in charge of data preparation, analytics, and applications. For this new role, controllers need expertise in advanced analytical techniques, data visualizations, business intelligence techniques, statistical and analytical models, and programming skills, and they need to be able to clean and organize large amounts of data (McAfee & Brynjolfsson, 2012).
Elbashir et al. (2013) argue that despite the significant investments of companies in business intelligence initiatives to support their operational and strategic planning, control, and decision-making processes, executives are dissatisfied with the information that the systems generate. Despite the vast amounts of data stored in the system ready to be used to support the business, the executives believe that the system gives a poor insight. Due to the lack of ability of employees to use data from a variety of sources, business intelligence failed to assist businesses’ executive decision-making at strategic as well as operational levels, as a result of which they failed to increase the value of a company. In addition, Graham et al. (2012) found that the technical adjustments in financial departments did not necessarily benefit experts.
Some Finance professionals expressed a sense of constraint in their department, citing a need
for more technical skills but a lack of resources to help personnel enhance their technical abilities.
Furthermore, the work conducted by Oesterreich and Teuteberg (2019) concludes that there exists a skills gap comparing the controller’s new skills requirements and the competence supplies evident in their skills profile. Even though business analytics has lately become a substantial element of the controller’s skill set, the findings of their study show that present controlling professional skill profiles do not meet the standards of the future controller.
However, this finding varies among companies as the consequences of their study for businesses will be determined by the strategic path that corporations choose to pursue in terms of the use of data analytics and the structure of the organization. In this regard, their study shows that as a firm grows larger, the business analytics and IT abilities in controlling professional profiles decreases. Companies must strive to narrow the skills gap if the controller’s jobs are truly extended by the data analytics activities. Organizations in a variety of industries are required to invest heavily in the employees' technical skills to stay competitive.
On top of that, it is widely understood that firms must provide training opportunities for people working at their firm to obtain adequate business analytics skill sets (Brands & Holtzblatt, 2015).
Matt et al. (2015) observed that digital innovations come together with the need for changing technical skills. They state that while present employees might not be technically minded and lack the technical skills to deal with forthcoming developments, it is also hard to find new employees who do meet the new requirements depending on the specific department.
According to their study, new research could contribute by showing whether improving technology capabilities would be helpful for businesses. This thesis will contribute to the literature by analyzing whether investments in the employees' skills in combination with digitalization, will lead to higher effectiveness of the Finance function.
There is a concern that the number of FTEs needed in the Finance function will decrease due to digitalization in the future (Acemoglu & Restrepo, 2020; Ford, 2015). However, this should not be a consideration if companies want an effective and influential Finance function.
According to the research of Frey and Osborne (2017), many decision-making positions will benefit from unbiased algorithmic solutions and this will provide controllers with new opportunities and responsibilities. On the one hand, algorithmic suggestions may be used as inputs to humans in the most difficult situations, and on the other hand, algorithms can be held accountable for making acceptable decisions on their own. For example, automated decision- making is used in the financial sector as systems can scan more financial statements, news
releases, and other data compared to any human, and then respond more quickly as a result of it.
Technology and data analytics are seen as transformative elements in organizations by management. Therefore, organizations deploy technologies to enhance reporting and decision- making. Controllers served as a company’s major source of information for decision-making and control. As a result, controllers have apparent ties to business intelligence technologies and may profit from their use. Unfortunately, according to studies of the literature in leading accounting and information papers, little study has been done on this connection thus far (Rikhardsson & Yigitbasioglu, 2018). Therefore, this thesis responds to calls from literature to investigate the effects of digitalization on the Finance function (Möller et al., 2020).
To get a more in-depth view of the effect of investing in the Finance professionals’
technical skills on the relationship between digitalization in Finance on the one hand and the effectiveness of the Finance function, on the other hand, survey research will be conducted among Finance professionals. As relatively little research has focussed on this theme, the insights obtained in this thesis will contribute to the literature. In addition, the practical application of this thesis will be examined by assessing whether the influence of both technologies and investing in the technical skills of Finance professionals are helpful for Finance-related business purposes.
1.2 Research question
Due to the increasing amount of data coming available it already has become unmanageable for humans to process all this information, necessitating the need for big data analytic techniques to help humans in making decisions. Despite the information from the research above, there is still no clear understanding of how the use of digital technologies leads to a more effective Finance function. That is why this thesis will investigate how the use of digital technologies affects operations in the Finance function of a company. In addition, the impact of investments in the technical skills of employees on this relationship will be examined. The following research question will be investigated:
“Does investing in the technical skills of financial employees affect the relationship between digitalization in Finance and the effectiveness of the Finance function?”
1.3 Reading guide
The thesis contains five chapters and the remainder is structured as follows. This introduction is followed by a review of the literature in chapter two in which the literature that this thesis is based on is reviewed in greater depth. Furthermore, the research hypotheses that are relevant to this thesis are presented. Chapter three concentrates on the empirical investigation, defining data, and measuring variables. After that, the results of the analysis are shown. The thesis finishes with a discussion of the results, and a variety of research and practice implications.
2 Literature review
2.1 Absorptive capacity
Cohen and Levinthal (1990) developed a theory where prior understanding of the subject helps to identify, assimilate as well as use the new information’s worth for economic gain. They suggest that one’s capacity to assess and use external information is substantially determined by past relevant knowledge. This background information might range from fundamental abilities to a common language, but it can also encompass knowledge of the latest scientific discoveries or technological breakthroughs in a certain area. Thus, prior relevant knowledge impacts the capacity to perceive new information’s value, digest it, and use it for business purposes. The sum of these abilities is referred to as a firm’s ‘absorptive capacity’ (AC).
The theory’s cognitive foundation for one’s AC is based on past relevant information and diverse background. The argument that the capacity to integrate information is a result of the variety of the prior knowledge system is based on two related ideas, namely that learning is a cumulative process, and when the learning topic is connected to what has previously been established, learning performance is at its peak. Due to this, learning is more challenging in new areas as well as, more broadly, the knowledge of an individual will only alter in little steps.
Furthermore, knowledge variety has an essential function. A broad background gives a more solid foundation for learning because, in a situation where the knowledge areas, from which potentially helpful information can arise, are unknown, it improves the likelihood that the new knowledge will link to what is already understood. Knowledge variety not only boosts assimilative skills but also aids the creative process by allowing one to form new connections and associations.
Thus far, the growth of a firm’s AC will be built on earlier investments in the growth of an individual’s AC and continue to expand cumulatively. However, the AC of a company is not merely the total of its workers’ AC, because AC does not only relate to a firm’s ability to absorb knowledge but also to the ability of the firm to take advantage of it. As a result, a firm’s AC is not only determined by its direct interaction with the outside world. It also relies on information transferred between and within departments that may be far distant from the site of introduction.
Furthermore, according to Van den Bosch et al. (1999), the features of an organization’s AC are related to the nature of expertise in its surroundings. They extended AC by looking at the influence of environmental conditions on the relationship between the absorptive capacity
process, assimilation and exploitation, and the outcomes. According to Cohen and Levinthal (1990) the bigger the firm’s competence and AC, the higher the chance to be open to new technical possibilities, and the more probable the aspiration level of the organization will be defined in terms of technological opportunities rather than based on performance indicators.
As a result, firms with greater AC will be more proactive, making use of chances that exist in the environment, regardless of existing performance. Firms with lower AC, on the other hand, will likely be reactive, looking for new options in reaction to failure on a non-technical performance criterion.
Past studies have linked AC to a variety of performance outcomes, including innovation, knowledge transfer, and financial success. A specific study done by Zou et al.
(2018) shows that the firm’s AC is positively related to its innovation as well as to its knowledge transfer performance. However, as changes within a firm come from top-level, operational-level employees within departments are the ones who are obtaining new expertise and executing the adjustments that facilitate the transition to increase performance. Yet, little research has been done on applying AC at the department level (Lin et al., 2016;
Riemenschneider et al., 2010). That is why the organizational view of AC is translated into the AC of the Finance function within an organization. This study applies AC in a technological context where AC of the Finance function indicates whether companies have adopted and integrated digital technologies. The exploitation of digital technologies in the Finance function could have a positive influence on the effectiveness of the Finance function. An effective Finance function is on the one hand focused on innovating the needs of the Finance function, and on the other hand exchanging financial knowledge to improve performance which should result in a more effective Finance function.
2.2 Effectiveness of the Finance function
The Finance function is one of the most important staff departments of any organization. In general, all financial processes in a company are referred to as the Finance function. The financial experts are normally found in the Finance department, although they may also be found in other sectors of the company. Finance functions operate at the management team level and have lines going across the business. A Finance function also has ties with all other organizational units. By maintaining the highest possible level of quality in its personnel, processes, and systems, the Finance function assists the company by maximizing the use of its resources (de Waal & Bilstra, 2016).
De Waal et al. (2019) suggest how the Finance function should evolve to achieve a high-performing status. One of these suggestions states that the focus of the Finance function needs to be shifted from ‘going concern’ to ‘continual improvement’. This means that the Finance function will be dedicated to enhancing its efficiency and effectiveness constantly.
Companies are in a perpetual state of transition, with change being the norm and affecting financial operations. In the future, the Finance function will prepare for these changes by anticipating them and ensuring that its procedures and services are up to date.
The Finance function of today is not only cash and capital management or profit planning and management anymore. The responsibilities of the Finance function have become wider, such as creating models to lead the search for new information, establishing performance targets, developing operating controls, and rationalizing operating standards (Chang et al., 2014). McAfee and Brynjolfsson (2012) argue that tasks of the Finance professionals nowadays also exist out of data preparation, analytics, and applications. The Dutch professional organization of accountants and the association of registered controllers state that the Finance professional works in four core areas, namely strategic management, performance management, Finance operations & reporting, and governance, risk & compliance (de Waal &
Bilstra, 2016). Finance operations & reporting, as well as governance, risk & compliance, serve as the Finance function’s foundation. Finance professionals supply the business with accurate and timely data in these two areas. The most essential goals here are to maintain continuity and to regulate the organization. The rising relevance of decision support by the Finance function is highlighted by strategic management and performance management. It is all about providing management with information and advice.
The research of Verstegen et al. (2007) complements this by saying that the responsibilities of controllers can be divided into two different groups. They say that on the one hand controllers are in charge of assisting in the making of business decisions. On the other hand, controllers are known as ‘corporate policemen’ as they are delivering information concerning the operations of the business unit by having a greater degree of supervision and maintenance of accounting systems and risk control. How effective a Finance professional is in performing their duties depends on the resources shifted towards the specific tasks (Chang et al., 2014).
2.3 Digitalization in Finance
The responsibilities of the Finance function go together with the possibilities that digital technologies offer nowadays. The adoption and use of digital technologies in the Finance
function are affected by the AC of the Finance function. This study will look at the relationship between the exploitation of digital technologies and the effectiveness of the Finance function.
Digitalization is the process of converting a business into a digital one. This will be done by the application of digital technology to alter a company model to create new revenue and value-generating possibilities (Glossary Gartner, 2022). In the financial context, digital technologies can on the one hand be used efficiently by automating routine tasks of Finance professionals (Ng et al., 2021). On the other hand, cognitive systems can interpret unstructured data and transform it into meaningful information, which in turn assists humans in decision- making (Kelly, 2015).
Intelligent automation, which is a combination of robotic process automation (RPA), artificial intelligence, and soft computing, assists firms in increasing the efficiency of processes, risk management, and conformity to quality and regulations (Huang & Vasarhelyi, 2019). RPA is software that follows rules and algorithms without the need for human intervention. The robot does repetitive activities, runs programs, and processes data, but considerably quicker and more reliably (Uskenbayeva et al., 2019). Routine and repetitive financial tasks may be automated to save money and release human resources for more essential tasks. In a digital system, RPA makes it possible to automate regular decisions and removes operational errors, consisting of well-structured, rule-based, and repeatable judgments employing vast amounts of data. The process complexity must be kept to a minimum and it should not include cognitive skills in the business process. Sales process assistance, systems for making schedules, assistance desk, and contact center operations are all typical instances of RPA. This is since these jobs are structured, and the logic of decisions is usually rules-based and repetitious with prescribed guidelines. RPA provides benefits to businesses by automating a portion of the business processes and releasing additional labor from dull and repeating tasks (Ng et al., 2021). Time spent on accounting systems, or on transaction processes, which can be classified as repetitive, low-judgment accounting tasks, is reduced by RPA (Graham et al., 2012). It has been widely used in the accounting industry for doing repetitive and low-judgment audit tasks (Huang & Vasarhelyi, 2019). A specific example that contributes to data-driven operational efficiency with the use of RPA in the financial industry is that it can support customer advisors, who work at banks. When they are having a conversation with a customer, the robot gets all of the customers' data from the call and displays it in one summary. The robot completes scripts 5,5 times faster than one employee, which boosts efficiency and response time (Lewicki et al., 2019).
RPA has progressed to a new level thanks to the emergence of artificial intelligence. In real-world scenarios, RPA should work in concert with other technologies. Cognitive activities that are automated are difficult to complete, yet not conducive to typical methods of automation. Artificial intelligence and machine learning are revolutionizing the industry and commercial processes by bringing intelligence to them (Zheng et al., 2018). Artificial intelligence allows decision-makers to observe, analyze, and adjust to their surroundings. More complicated tasks, including judgemental action in the business procedure, as well as human perceptual capabilities, are ideal for intelligent automation using fuzzy logic, machine learning, and natural language processing. As a result, intelligent automation becomes a more useful tool (Coombs et al., 2020).
Next to RPA, there are three other types of intelligent automation, namely Intelligent Process Automation (IPA), Augmented Intelligent Process Automation (AIPA), and Autonomous Agents (AA). IPA, which combines RPA, artificial intelligence, and cognitive skills, uses unstructured data such as images as well as structured data to execute prescriptive analytics. With the use of artificial intelligence and soft computing, IPA can give a certain degree of cognitive choice and can replicate human decisions. Since the decision-making process is not based on rules, just a minimal amount of human interaction is required (Ng et al., 2021). AIPA enables a comprehensive approach to corporate process automation and digital operation by advancing to the next stage of cognitive decision quality. Decision engines must be included with fast judgment and deductive analytics, and with the cognitive capacity comparable to human intellect in AIPA (Yin et al., 2020). AA are software programs that can react to the situation and make requests on their own, without the need for human interaction.
The key aspects of AA are excellent handling abilities, attaining a high level of decision intellect with no human interaction and the ability to self-learn, quality assurance, and exception handling decision assistance. Knowledge augmentation helps the system to make superior decisions as human intellect (Ng et al., 2021).
Further examples of intelligent automation in Finance where cognitive technologies with skills that are similar to those of humans are mentioned in the research of Pramanik et al.
(2019). It is used to provide predictive advice within Investment Banks, but also for other cases ranging from identifying abnormalities for fraud and cyber-security to the creation of specific trading strategies to advise customers. Furthermore, banks are using virtual chatbots for interacting with customers without human intervention. Lewicki et al. (2019) mention a recent development where it is possible to generate yearly reporting, pay stubs, income statements, and invoicing.
There are several digital technologies and techniques which support decision-making, such as big data analytics and machine learning. As data gathering becomes easier and less expensive, professionals who can work with these big amounts of data are becoming more valuable (McAfee & Brynjolfsson, 2012). The utilization of big volumes, high velocity, and a variety of data sources in the decision-making process distinguish analytics from traditional decision support technology (McAfee & Brynjolfsson, 2012). Furthermore, machine learning models, e.g. deep learning, can be useful in Finance because of the large amounts of data, high velocity, and the involvement of patterns. It is used to identify what items to offer clients or to determine their retention probability in consumer banking. Deep learning models also provide opportunities to gain an understanding of how the market values options (Culkin & Das, 2017).
Machine learning algorithms can also be used to detect credit-card fraud in online retail (Carneiro et al., 2017).
According to the theory of Cohen and Levinthal (1990), employees with prior related expertise are better able to assess the value of new information, assimilate this information and exploit this to increase performance. As digitalization in the Finance function is upcoming, more and more companies are looking for ways to digitalize their Finance processes. When Finance professionals use digital technologies such as process automation, data visualization, or advanced analytical techniques, it indicates that the Finance function has adopted and integrated digital technologies in its activities and operations. The more tools they exploit, the higher the AC of the Finance function is. As prior studies have shown that AC is positively related to its innovative as well as knowledge transfer performance, the focus of AC of the Finance function in this study captures whether exploiting digital technologies enhances the effectiveness of the Finance function.
2.4 Technical skills
As with the rise of new technologies used in the Finance function, new skill sets of employees are required. These are not only required for the change but also for routine operations after that. Existing employees may have a distinct, less technical perspective and lack the technical skills needed to deal with future changes, but new highly trained, and focused employees may be hard to find (Matt et al., 2015). The ability to use new types of technology to get wisdom from large amounts of data requires technical skills. Most big data investments are a waste of money since the majority of businesses are either not prepared or do not make judgments based on the information gleaned through data (Gupta & George, 2016). Some examples of these
skills are knowing how to use intelligent automation, machine learning, data visualization, statistical and analytical models, and understanding programming languages (McAfee &
Brynjolfsson, 2012). Companies weaken themselves by failing to acquire technological skills for the operational workforce, probably this will also hold for the Finance workforce. As a result, activities are poorly coordinated and the organization’s financial performance is continually hampered (Gupta et al., 2016). It is not just the technology that secures the intended business outcome, but also the expertise. However, difficulties arise with finding new talent and training the existing workforce to increase their technical abilities, because dealing with big data necessitates new technological skills that aren’t typically taught in school. Some universities have begun to teach courses in these skills, but there is still a considerable scarcity of people with technical skills in big data (Gupta & George, 2016); this is also likely to be the case in Finance. Furthermore, companies themselves must provide opportunities for Finance professionals to develop their technical skills. Continuing education is required due to the frequent developments in business analytics. Finance professionals can obtain information through internal or external training, or by working with consultants. Other training tools include case studies created by software providers and professional groups and on-demand training videos (Brands & Holtzblatt, 2015).
2.5 Hypothesis development
As a result of recent technological advancements and regulatory reporting requirements, Finance professionals have now access to tools to properly structure data and the necessary instruments to assess this data. This research investigated whether Finance professionals understand how to utilize the tools to achieve a more effective Finance function. Below, hypotheses are provided based on the theory and previous research presented in the preceding paragraphs.
The work conducted by Jarrahi (2018) states that humans and artificial intelligence have complementary roles in the context of organizational decision-making. They have distinct but complementary abilities required for effective decision-making. Furthermore, according to Martínez-Caro et al. (2020), business digitization results in the re-engineering of business processes which leads to, among others, improved decision quality, business improvements, and enhanced organizational effectiveness as an outcome. Schneider et al. (2015) state that data analytics provides tools to investigate data from financial transactions in three ways, namely to infer, forecast, and assure activities. This may be done by utilizing data analytics to deduce
successful cost-cutting methods and prospective workflow improvements from corporate expenditure trends. Data analytics can also assist financials in effectively making decisions in a business by predicting future stock performance or as continuous monitoring.
However, the research of Byrne and Pierce (2007) found that the degree to which technology assists in lowering the amount of time needed to complete tasks, resulting in a more effective Finance department, varied. Some financial managers even said that the introduction of new technologies did not lead to benefits for the Finance professional at all. In addition, Tang et al. (2014) investigated how visualization and interactivity of financial data affect financial decision-making performance. These two aspects have been highlighted as potentially improving financial statement analysis. Their research shows that when visualization and interactivity are supplied separately, financial decision-makers are less calibrated. So, visualization in itself leads to less accuracy in financial decision-making, which could result in making poor decisions and reducing the effectiveness of the Finance function.
Building on AC theory, the appropriate use of digital technologies by Finance professionals makes it possible to carry out their responsibilities more effectively. Finance professionals can use tools for automating routine processes, standardization, analyzing data, predicting future performance, and so on. Automation makes it possible to reduce time spent on routine activities, which can be invested in value-adding activities. Data analytics allows for greater insights to be provided based on available data. The more digital tools the Finance professionals have adopted and integrated to exploit data, the higher the AC of the Finance function. When Finance professionals exploit digital technologies to a high degree, we expect that this enhances the effectiveness of Finance professionals performing their responsibilities.
Based on the aforementioned literature, the following hypothesis is formulated:
HI: Digitalization in Finance is positively related to the effectiveness of the Finance function
Within the scientific community, there is some controversy about whether the Finance function needs to improve its skills to perform data analytic techniques or leave these activities to a data scientist (Mödritscher & Wall, 2017; Payne, 2014).
Payne (2014) states that data analytic and information technology skills have become more important as part of the Finance professional skills profile. These technical skills are
needed to perform the responsibilities of the Finance function, nowadays (Gupta et al., 2020).
Hiring well-skilled technical Finance professionals will increase the technical capabilities of the Finance function. The existing workforce can learn from the well-skilled Finance professionals and will adapt new skills that will allow them to perform their responsibilities as Finance professionals more effectively (Gupta & George, 2016). The research of Kusi-Sarpong et al. (2021) states that a lack of education and training in relevant technologies and procedures prevents employees from possessing the recent abilities necessary to process the new tools.
However, when they do get these training opportunities employees will be more motivated to use the new tools. Gupta et al. (2016) state that companies get weaker as a result of their failure to acquire technical capabilities for their employees. As a result, the activities of the Financial professionals are disorganized, and the effectiveness of the Finance function will be lower.
Given the previous literature review, it is expected that when companies invest in the technical skills of financial employees, they will do their responsibilities in a more effective way which will lead to a more effective Finance function.
This leads to the second hypothesis:
HII: Investment in the technical skills of Finance professionals will lead to a more effective Finance function
According to Oesterreich and Teuteberg (2019), business analytics capabilities and systematic and statistical competencies have become significant for the Finance function.
However, they argue that a skills gap in the expertise of digital technologies exists when comparing the new skills demands of the controller to the skills that most controllers have to offer. In addition, Krumwiede (2017) conducted a survey among Finance professionals where the professionals believe they have a comprehensive grasp of artificial intelligence, but lack the expertise to offer value to artificial intelligence tasks.
Based on AC, the use of digital technologies in Finance has a positive effect on the effectiveness of the Finance function. However, when the Finance function does not have the right expertise and background in the digital technologies that they use, they are not able to exploit the tools sufficiently, and therefore it does not necessarily lead to well-coordinated activities, improved financial decision-making, or achieving better financial results (Gupta et al., 2020). Therefore, the knowledge environment of the employees should change by investing
in the technical skills of the employees, as suggested by the research of Van den Bosch et al.
(1999). Investment in the technical skills of Finance professionals is used as a moderator to see whether this strengthens the relationship between digitalization and the effectiveness of the Finance function. This leads to the third hypothesis:
HIII: Digitalization in Finance, together with investing in the technical skills of Finance professionals will lead to higher effectiveness of the Finance function
The conceptual model of the hypotheses is illustrated in figure 1:
Figure 1: Conceptual model
Digitalization in Finance
Effectiveness of the Finance function Investment in
technical skills
HI
HII HIII
3 Research methodology
3.1 Research design
To answer the research question and test the hypotheses I join the thesis survey project of the University of Amsterdam led by dr. Verbeeten. The project is about digitalization in the Finance function. To join the project and get access to the database ten suitable participants, which are checked by dr. Verbeeten, must submit the survey. An example of a suitable respondent is someone with some working experience in the Finance function and who is still working as a Finance professional in a Finance department. The criteria guarantee that the data from the survey is accurate and valuable in answering the research question. The advantage of joining this project is that a sufficient amount of respondents for the survey is guaranteed. In addition, with the constructs of the survey, I will be able to operationalize the hypotheses and answer the research question.
3.2 Sample characteristics
As not all respondents are suitable for this thesis some answers have been excluded. The sample used in this thesis includes 182 observations while there were 192 respondents. As this thesis only focuses on respondents from Europe, ten observations from China have been removed from the data set. The respondents all have a job in Finance, wherefrom 31 are CFO, 9 are group controllers, 44 are Finance managers, 43 are business controllers, 26 are financial controllers and the other 29 have different jobs in Finance. There are 37 single-business organizations, 73 multi-divisional organizations where the respondent works in Finance at the corporate level, and 72 multi-divisional organizations where the respondent works at the business unit level. The firm size of the entities where the respondents work is given based on the total sales of the entity in 2020 (in € mln): 109 of the respondents work in a company with sales <250, 37 of the respondents work in a company with sales between 250 and 1.000, and 36 of the respondents work in a company with sales >1.000. The sample has been split up into four industries, where 70 of the respondents work in the manufacturing sector, 23 in the financial sector, 60 in the non-financial service sector, and 29 in the public sector.
3.3 Variable measurement
3.3.1 Dependent variable
The dependent variable in this research is the effectiveness of the Finance function. The construct to measure the effectiveness of the Finance function is based on prior research by Chang et al. (2014). Respondents were asked how effective on a five-point Likert scale (1 = not effective at all, 5 = extremely effective) the Finance function of their entity is in achieving the following: (1) providing relevant information to business decision-makers, (2) measuring and monitoring business performance, (3) continuous business improvement, (4) aligning Finance with business requirements, (5) driving cost reduction and efficiencies in the business, and (6) providing inputs for identifying and executing growth. Factor analysis has been carried out to determine the extent to which the tasks of the Finance function rely on a single factor.
The component analysis revealed that the six questions are all loaded on the same factor, appendix A contains the results. Furthermore, to assess the factor’s reliability, an examination of Cronbach Alpha has been performed. According to Van Griethuijsen et al. (2015), a Cronbach Alpha is acceptable when the values are higher than 0,6 or 0,7. There is a Cronbach Alpha of 0,906 which means that the results are reliable. Based on the findings of the factor analysis and the Cronbach Alpha, the six questions have been integrated into one component.
This has been done by taking the average of the six questions, leading to the new variable EFF_FIN. The average has been taken which makes it possible to compare the results with other studies.
To ensure the validity of the dependent variable a check has been done on the correlation between the effectiveness of the Finance function and how important the following objectives for the Finance function of the respondents' entity are (1) providing insights and advice to business managers, (2) influencing operational decision-making, (3) developing strategic objectives and business models, (4) aligning activities across the organization to drive value generation, and (5) driving major changes in the organization. The respondents could answer the questions on a five-point Likert scale (1 = not at all important, 5 = extremely important). It is expected that there is a positive correlation between the Finance function’s objectives and the effectiveness of the Finance function. A Pearson correlation test has been done to determine the degree of correlation between the EFF_FIN and the five objectives of the Finance function. As shown in table 1 a positive and significant correlation exists between EFF_FIN and the five objectives (p < 0,01) which means that EFF_FIN is a valid variable.
Table 1 Pearson coefficients for validity check EFF_FIN
3.3.2 Independent variable
The independent variable in this research is digitalization in Finance. The measure for digitalization in Finance is based upon the research of Srinivasan and Swink (2018) and Ritter and Pedersen (2020). For the independent variable, Finance professionals have answered eight questions on a five-point Likert scale. The respondents were asked to indicate to what extent the following statements apply to the work of the Finance function in their entity: (1) we deploy tools (e.g. robotic process automation software) for automating routine processes, (2) we have installed tools to eliminate manual work for repetitive activities, (3) we use digital technologies to standardize processes, (4) there is a high level of automation and standardization, (5) we use advanced analytical techniques (e.g. optimizations, simulation, regression) to predict future performance (e.g. sales forecasts, profitability scenarios), (6) we develop and use statistical and analytical models to inform decision-making, (7) we analyze large amounts of data (e.g. from customers, suppliers, competitors) with the help of analytical tools and models, and (8) we use cognitive computing techniques (e.g. artificial intelligence, machine learning, deep learning, neural networks) to improve decision-making. The response scale ranges from 1= not at all to 5 = to a very high extent.
As some of the digital technologies can be used more efficiently, and other technologies are translating knowledge into meaningful information to assist humans, factor analysis has been performed to see how much one technology weighs on the related technologies. The Principal Component Analysis revealed that the nine questions are all loaded on the same factor with an eigen value of higher than 1. However, the scree plot indicates that there may be two factors because the second factor is close to the eigen value of 1. Later in this thesis, the two factors will be checked, therefore question five is excluded, because when there are two factors created, five loaded as much on the one as on the other. Therefore, the remaining eight questions are used as one factor, appendix B contains the results. To assess the factor’s reliability, an examination of Cronbach Alpha has been performed. There is a Cronbach Alpha of 0,918 which means that the results are reliable. Based on the findings of the factor analysis
Pearson Correlation
(1) (2) (3) (4) (5)
0,449** 0,461** 0,474** 0,376** 0,257**
** p<0,01, * p<0,05
Variables EFF_FIN
and the Cronbach Alpha, the eight questions have been integrated into one factor. This has been done by taking the average of the eight questions and leading to the new variable DIG_FIN.
To ensure the validity of the independent variable a check has been done on the correlation between digitalization in Finance and two purposes wherefore digital technologies would be used in the Finance function. The two purposes are (1) budgeting and target setting, and (2) short-term planning and forecasting. The respondents could rate the purposes on a six- point Likert scale (1 = not applicable, 6 = to a very high extent applicable). It is expected that there is a positive relationship between the two purposes and digitalization as digital technologies would be used for the two purposes. The performed Pearson correlation test, in table 2, shows a positive significant relationship (p < 0,01) between digitalization and the two purposes which means that DIG_FIN is a valid variable.
Table 2 Pearson coefficients for validity check DIG_FIN
The moderator used on the relationship between the use of digital technologies and the effectiveness of the Finance function and the variable that is used for the direct effect on the effectiveness of the Finance function is the investment in the employees' technical skills. The components to measure the independent variable are based on prior research by Chakravarty et al. (2013), Gupta and George (2016), Gupta et al. (2020), Eller et al. (2020), and Wang et al.
(2019). Respondents were asked to indicate to what extent the Finance function of their entity (1) provides the resources and opportunities for employees to obtain data analytic skills, (2) hires new employees that already have data analytic skills, and (3) has the right skills and expertise to take advantage of data analytics. The respondents could answer on a five-point Likert scale where1 = not at all and 5 = to a very high extent.
A factor analysis was performed on the technical skills of the Finance function to identify the extent to which the questions load on one factor. The Principal Component Analysis revealed that the three questions are all loaded on the same factor, appendix C contains the results. To assess the factor’s reliability, an examination of Cronbach Alpha has been performed. There is a Cronbach Alpha of 0,819, to be found in appendix C, which means that the results are reliable. Based on the findings of the factor analysis and the Cronbach Alpha, the three questions have been integrated into one factor. This has been done by taking the average of the three questions and leading to the new variable INV_SKILL.
Pearson Correlation
(1) (2)
0,498** 0,545**
** p<0,01, * p<0,05
Variables DIG_FIN
A correlation check has been performed to evaluate the validity of the moderator. A Pearson correlation test has been performed between the investment in the employees’
technical skills of the Finance function and the investments in digital technologies of the company. It is expected that when the organization invests strongly in digital technologies, they also invest to a high degree in the skills of the employees to use these digital technologies. The following two statements are used to measure the investment in digital technologies of the company, (1) we invest a lot in new digital technologies to support various organizational processes, and (2) we have a climate that is supportive of using digital technologies.
Respondents could indicate to what extent they agree with the statements on a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree). The Pearson correlation test shows a positive and significant correlation (p < 0.01), as shown in table 3, which means that INV_SKILL is a valid variable.
Table 3 Pearson coefficients for validity check INV_SKILL
3.3.3 Control variables
The comprehensive design of the survey offers possibilities to control for other factors that may affect the effectiveness of Finance. The following control variables are taken into account:
industry, firm size, and the degree of digitization in a company.
Firstly, to measure the control variable sector a construct of the survey is used where the respondent had to answer the question about the entity’s primary industry (by revenues).
The respondent had 17 options to choose from, for this thesis the nineteen industries are divided into four categories, namely (1) manufacturing (agriculture, hunting, forestry and fishing, mining and quarrying, traditional manufacturing, high-tech manufacturing, utilities, construction, and building), (2) non-financial service (wholesale and retail trade, transportation and storage, accommodation and food service activities, information and communication, real estate activities, professional, scientific, and technical activities) (3) financial sector (financial and insurance activities), and (4) public sector (public government, defense organizations and social security, education, human health, and social work, environment, culture, recreation, and other service activities).
Pearson Correlation
(1) (2)
0,505** 0,445**
** p<0,01, * p<0,05
INV_SKILL Variables
Secondly, to measure the control variable firm size a construct of the survey is used where the respondents had to indicate the total sales of the entity in 2020 (in € mln). The nine options the respondents could choose from were a total sales of the entity in 2020 (in € mln) of (1) < 10, (2) 10 – 49, (3) 50 – 99, (4) 100 – 249, (5) 250 – 499, (6) 500 – 999, (7) 1.000 – 4.999, (8) 5.000 – 15.000, and (9) > 15.000.
Thirdly, the degree of digitization of the organization has been taken into account as a control variable. In the survey, the respondents could indicate to what extent they agreed on a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree) with the following statements about their entity (1) we have a clearly defined digital strategy, (2) we invest a lot in new digital technologies to support various organizational processes, (3) we have established new business models, products, or services based on our digital competences, and (4) we constantly seek new ways to enhance the effectiveness of digital technologies use. As this variable could overlap with the underlying constructs of DIG_FIN, factor analysis has been performed to check if they both rely on the same factor. As it turns out, the four elements do not rely on the same factor as the underlying constructs of DIG_FIN as shown in table 4.
Therefore, this control variable is valid to use.
Rotated Component Matrix
Component
1 2
(1) 0,783 0,309
(2) 0,862
(3) 0,878
(4) 0,836
DIG_FIN (1) 0,728
DIG_FIN (2) 0,792
DIG_FIN (3) 0,339 0,722
DIG_FIN (4) 0,433 0,699
DIG_FIN (5) 0,769
DIG_FIN (6) 0,820
DIG_FIN (7) 0,749
DIG_FIN (8) 0,810
Table 4 Factor analysis of degree of digitization
At last, the entity’s average annual growth in revenues over the last three years has been taken into account as a control variable. The respondents had five options to choose how much the entity’s average annual growth in revenue was over the last three years (in %), namely (1) < 0%, (2) 0 – 10%, (3) 11 – 20%, (4) 21 – 30%, and (5) >30%.
3.3.4 Summary variables
The table below, table 5, provides a summary of the variables that will be used in the remainder of this thesis.
Table 5 Summary of the variables
Name Description Type of variable
EFF_FIN Effectiveness of the Finance function Dependent variable
DIG_FIN Digitalization in Finance Independent variable
INV_SKILL Investment in the employees' technical skills Independent variable
D_MANUFACTURING Sector: manufacturing Control variable
D_FSERVICES Sector: financial services Control variable
D_NFSERVICES Sector: non-financial services Control variable
D_PUBPLICSEC Sector: public sector Control variable
FIRM_SIZE Firm size Control variable
DEG_DIG Degree of digitization of the organization Control variable
ANN_GROW Average annual growth in revenue Control variable
4 Results
4.1 Descriptives
The descriptives for the variables used in this thesis are discussed in more detail in this section.
Table 6 provides the descriptive statistics.
Descriptive Statistics
Variables N Mean Std. Dev. Min Max
EFF_FIN 182 3,157 0,747 1 5
DIG_FIN 182 2,416 0,885 1 5
INV_SKILLS 182 2,775 0,872 1 4,3
D_MANUFACTURING 182 0,385 0,488 0 1
D_FSERVICES 182 0,126 0,333 0 1
D_NFSERVICES 182 0,330 0,471 0 1
D_PUBLICSEC 182 0,159 0,367 0 1
FIRM_SIZE 182 4,080 2,401 1 9
DEG_DIG 182 4,881 1,368 1 7
ANN_GROW 182 2,450 1,090 1 5
Table 6 Descriptive Statistics
Table 6 shows the total of observations (N), mean, standard deviation (Std. Deviation), minimum, and maximum values for each variable. Furthermore, for the dependent variable effectiveness of the Finance function (EFF_FIN) is the minimum 1 and the maximum 5, whereby 1 means that the Finance function is not effective at all and 5 means that the Finance function is extremely effective. The mean of EFF_FIN is 3,16 which is around the average of 3. Furthermore, there is sufficient variation in the answers as the standard deviation is 0,75.
The independent variable DIG_FIN has a mean of 2,42 which is a bit lower than the average of 3 and a standard deviation of 0,88 which means that there is sufficient diversity in the answers. INV_SKILL shows a mean of 2,77 which is close to the average of 2,67. The standard deviation of 0,87 shows that there is enough variation in the answers.
Most of the observations are from the manufacturing sector (38,5%) followed by the non- financial services sector (33%). Only 15,9% of the respondents work in the public sector, and 12,6% in the financial service sector. The control variables firm size, degree of digitization, and average annual growth have relatively a higher standard deviation. Firm size has a standard deviation of 2,401, degree of digitization of the organization has a standard deviation of 1,368 and average annual growth has a standard deviation of 1,090. The standard deviation shows how far the answers can deviate from the mean. So, with these three control variables, the
answers of the respondents are more spread out over the entire range. The descriptive statistics indicate that the dependent, as well as the independent and control variables, have sufficient spread for further statistical analysis.
4.2 Correlations
In this paragraph, the correlations between the variables are described. To check the correlations between the dependent and independent variables the Pearson Correlation test is used. The results of the test are given in table 7. Significant correlations have a p-value of
<0,05. Next to the Pearson correlation test has the Spearman correlation test been performed.
The results of the Spearman correlation test, which can be found in Appendix D, are in line with the Pearson correlation test.
Table 7 Pearson Correlation test. For further details on the variables refer to table 5.
Pearson Correlation (1)(2)(3)(4)(5)(6)(7)(8)(9)(10) (1)EFF_FIN1,00 (2)DIG_FIN0,62**1,00 (3)INV_SKILLS0,63**0,68**1,00 (4)D_MANUFACTURING0,070,020,031,00 (5)D_FSERVICES0,110,040,07-0,30**1,00 (6)D_NFSERVICES0,070,070,03-0,55**-0,27**1,00 (7)D_PUBLICSEC-0,29**-0,15*-0,15*-0,34**-0,17*-0,31**1,00 (8)FIRM_SIZE0,16*0,25**0,19*0,09-0,03-0,03-0,061,00 (9)DEG_DIG0,48**0,57**0,60**-0,030,16*0,07-0,20**0,121,00 (10)ANN_GROW0,080,090,120,020,16*0,00-0,18*-0,060,131,00 ** p<0,01, * p<0,05
Variables
The results from table 7 show some significant correlations. DIG_FIN and EFF_FIN show a significant positive correlation of 0,62. This is in line with the expectations that there would be a positive relationship between digitalization in Finance and the effectiveness of the Finance function. Furthermore, INV_SKILLS shows two significant and positive correlations with the dependent and the independent variable. There is a positive and significant correlation of 0,63 with EFF_FIN, which is in line with the second hypothesis, and a positive and significant correlation of 0,68 with DIG_FIN. Organizations appear to invest in the digitalization of the Finance department as well as in the technological capabilities of its employees.
The results show a significant and negative correlation of -0,29 between the public sector and the effectiveness of the Finance function. This means the effectiveness of the Finance function is lower in the public sector. There is also a negative and significant correlation between the public sector and DIG_FIN (coef: -0.15) as well as the public sector with INV_SKILL (coef: -0,15). This implies that in the public sector organizations employ fewer digital technologies and invest less in the technical skills of employees.
Firm size shows a positive and significant correlation of 0,16 with EFF_FIN, 0,19 with INV_SKILL, and 0,25 with DIG_FIN. This means that the size of a firm has an impact on how effective the Finance function is, whether digital technologies are being used in Finance and if there is invested in the technical skills of employees. The coefficient of DIG_FIN is closest to 1 of the three significant correlations above which means that there is the strongest correlation between firm size and DIG_FIN. The size of the firm has the biggest impact on whether digital technologies are being used in Finance.
The degree of digitization of a company also shows three significant and positive correlations with EFF_FIN (coef: 0,48), DIG_FIN (coef: 0,57), and INV_SKILL (coef: 0,60).
This implies that the more digitalized a company is, the more effective the Finance function is, the more digital technologies in Finance will be used, and there will be invested more in the technical skills of employees.
4.3 Regression analysis
In this paragraph the regression analysis will be performed and the two hypothesis will be tested. This thesis considers two hypothesis whose regressions look as follows:
