Intellectual Capital and Performance of Banking and Financial Institutions in Panama: An Application of the VAIC™ Model

1. Introduction

Currently, intellectual capital (IC), together with financial capital, are considered key factors for the profitability of companies (Alipour 2012). In the new knowledge-based economy proposed by Drucker (1993), knowledge ceases to be just another resource alongside traditional factors of production, such as capital and land, and instead becomes the most significant resource. This shift has brought the importance of intangible resources to the forefront, giving them the recognition they deserve (Tiwari 2022). Consequently, IC is now projected as a crucial component in creating value for companies, contributing through sustainable competitive advantage(Awwad and Qtaishat 2023; Bombiak 2023; Jardon 2015; Shah et al. 2023; Teece et al. 1997; William et al. 2019; Xu et al. 2019). The management of intellectual assets has become essential for generating organisational value (Bontis 1998), with companies increasingly relying on intellectual potential over physical capital (Pulic 1998). In this new economy, intangible assets such as IC are fundamental (Özer and Çam 2016). However, measuring these intangible resources remains a challenge for public administrations (Abdulsalam et al. 2011) since IC continues to be a critical concept for analysing and reflecting the real value of organisations (Arslan and Kızıl 2019). Financial statements have limitations in capturing the full value of companies, based on the fact that the source of economic value has shifted from the production of material goods to the creation of IC (M. Chen et al. 2005). Recognising IC as an important business asset is essential, as it can generate sustainable competitive advantages and superior financial results (Barney 1991).

IC is defined as the possession of knowledge, application of experiences, organisational technology, customer relationships, and professional skills (Edvinsson 1997). According to Stewart (1997), IC is an intellectual material that has been formalised, captured, and leveraged to create wealth through the production of higher-value goods. The IC dimensions comprise three harmonising groups: people, structures, and relationships (Pulic 1998).These intangibles consist of human capital (HC), structural capital (SC), and relational capital (RC) (Edvinsson and Malone 1997; Stewart 1997).These structures are based on the Intellect Model (EUROFORUM 1998). Based on this model, in the Intellectus model HC refers to the knowledge, whether explicit or tacit, individual or social, that people and groups possess. Additionally, it considers their ability to generate benefits for companies (Bueno et al. 2011) through a combination of knowledge, skills, experiences, competencies, creativity, and individual capabilities (Bontis 1999; J. Chen et al. 2004; Edvinsson and Malone 1997; Sveiby 1997). Thus, employees are the carriers of knowledge, which is the crucial substance of products and services (Pulic 2008). SC encompasses the knowledge embedded in the company’s internal processes and technological systems, differentiating it from other organisations (Larios Prado 2009). RC can be defined as the knowledge resulting from relationships with market agents and society in general (Bueno et al. 2011).

To enhance the value generated by IC, various measurement models exist. In this study, the Value-Added Intellectual Coefficient (VAIC)™ model is utilised(Pulic 1998, 2000, 2004, 2008),which aims at measuring the value created within the organisation. The VAIC™ model calculates IC based on financial data using the following three components: Human Capital Efficiency Coefficient (HCE), Structural Capital Efficiency Coefficient (SCE), and Capital Employed Efficiency Coefficient (CEE). The VAIC™ indicates the overall state of corporate intellectual capacity, making it possible to measure the performance of intellectual potential and enable management intervention (Pulic 1998).The VAIC™ formula, which uses financials and accounting reports, is considered a suitable tool for measuring IC value creation (Bykova and Molodchik 2012; Pardo-Cueva et al. 2018; Pulic 2000; Shaban and Vijayasundaram 2019; Śledzik 2012; Sumedrea 2013; Villegas González et al. 2017).

The VAIC™ indicators are relevant, useful, and informative for stakeholders as they identify trends and link with traditional financial indicators (Demuner Flores et al. 2017). In addition, this model allows for comparative analysis between companies in the same competitive sector by using standards that measure their effectiveness. Thus, it provides information on company value, performance, and competitiveness, enabling the measurement of IC efficiency (Śledzik 2012; Pardo-Cueva et al. 2018).

Various studies have used the banking and financial sector as samples to measure IC through the VAIC™ model (Abdulsalam et al. 2011; Al-Musali and Ismail 2014; Chen Goh 2005; Demuner Flores et al. 2017; Duho 2020; Duho and Onumah 2019; Faruq et al. 2023; García Castro et al. 2021; Mavridis 2004; Meles et al. 2016; Mollah and Rouf 2022; Oppong and Pattanayak 2019; Ozkan et al. 2017; Singh et al. 2016; Soewarno and Tjahjadi 2020; Tran and Vo 2018).This study uses the VAIC™ model to analyse the relationship between return on equity (ROE) and the components of the VAIC™ since this allows for the quantifiable and objective measurement of IC’s contribution to companies, particularly the banking sector, which is classified as knowledge-intensive(Oppong and Pattanayak 2019).

This work contributes to the literature on IC by contrasting the results offered by the VAIC™ model with a resource-based view (RBV). This theory is based on the premise of heterogeneity among companies, suggesting that organisations differ in their resources and capabilities, with some resources being more valuable than others. According to RBS, a company's ability to generate resources depends on establishing competitive advantages. For resources to be considered competitive advantages, they must be scarce and difficult to imitate. IC is considered a key element in creating and maintaining these competitive advantages (Barney 1991;Bueno et al. 2011; Grant, 1991).

Companies in the banking and financial sectors in Panama listed on the Latin American Stock Exchange (LATINEX) between 2014 and 2020 were studied. These sectors were analysed because they represent an important pillar in the Panamanian economy, demonstrating their resilience during global financial crises, such as the COVID-19 pandemic, by maintaining profitability levels similar to those obtained in previous periods. In the banking sector, IC is considered a key factor for achieving competitiveness(Van Nguyen and Lu 2023) since banks must continuously innovate and remain competitive to survive. Due to their participation in the stock market, the companies analysed are required to publish their financial reports on their websites, providing reliable data (Arslan and Kızıl 2019).

This study is divided into five parts. After the introduction, the literature review and formulation of hypotheses are presented. The third part outlines the model and variables used in the empirical study, as well as the sample and the methods of data collection and processing. The fourth part presents the statistical analysis, evidencing the different assumptions that validate the model. Next, the results obtained are discussed. Finally, the main conclusions of the study, its limitations and future lines of research are presented.

2. Literature Review and Hypotheses Development

2.1. Literature Review

The following is a review of studies that utilised the VAIC™ model to measure the efficiency of the IC in the banking sector.

Arslan and Kızıl (2019): They applied the VAIC™ model to measure and compare the IC of Turkey's banks listed on the Borsa Istanbul Banking Index (BIST XBANK). Using Pearson's correlation test, they found a moderate relationship between ROE and VAIC™, with no relationship between ROE and indebtedness. A strong correlation was found between the HCE coefficient and the SCE and CEE variables, as well as a very close correlation between HCE and VAIC™. There was also a strong correlation between SCE and VAIC™ and a moderate correlation between CEE and VAIC™. The HCE, SCE, and CEE variables maintained a moderate correlation with the Return on Assets (ROA) variable. The HCE and CEE variables showed a moderate relationship with indebtedness, while a low correlation was found between SCE and indebtedness. Finally, a close correlation was observed between VAIC™ and ROE, a strong correlation between VAIC™ and SCE, a moderate relationship between VAIC™ and the ROE and CEE variables, and a low relationship between the VAIC™ and indebtedness.

Al-Musali and Ismail (2014): This study examined the IC performance of listed banks in Saudi Arabia between 2008 and 2010 using the VAIC™ model and investigated the impact of IC on financial performance. The results indicated a positive and significant association between VAIC™ and ROE, as well as between HCE and ROE. However, SCE showed negligible associations with financial performance indicators. A significant positive relationship was observed between the CEE and ROE. In addition, they analysed the SIZE variable; however, no relationship was found.

Meles et al. (2016):This study evaluated the efficiency of IC in a large sample of 5,749 commercial banks in the US from 2005-2012 using the VAIC™ model through the Ordinary Least Squares (OLS) method. They found a statistically significant relationship between VAIC™ and ROE. Moreover, they discovered a significant relationship between HCE and ROE, indicating that efficiency in using human capital (HC) has a greater impact on financial performance than other components of IC efficiency. They also found a significant relationship between asset SIZE and ROE.

Soewarno and Tjahjadi (2020): This study used multiple regression analysis to examine the relationship between IC and the financial performance of banks listed on the Indonesian Stock Exchange between 2012 and 2017. They found a statistically significant relationship between the SCE, CEE, and indebtedness variables and the ROE variable. However, no relationship was found between the HCE and SIZE variables with the ROA variable.

Mollah and Rouf (2022): This study analysed the impact of IC on the financial performance of all listed commercial banks in Bangladesh between 2014 and 2018, using the VAIC™ model. The results showed that HCE and CEE have statistically significant relationships with bank performance.

2.2. Hypotheses Development

Table 1 summarises the results of the literature review.

Based on these results, the following hypotheses are formulated:

H_a 

: There is a significant relationship between ROE and at least one of the variables VAIC, HCE, CEE, SCE, SIZE, and INDEBTEDNESS for banking and financial institutions listed on LATINEX between 2014 and 2020.

H₀ 

: There is no significant relationship between ROE and any of the variables VAIC, HCE, CEE, SCE, SIZE, and INDEBTEDNESS for banking and financial companies listed on LATINEX between 2014 and 2020.

3. Methodology

The study is based on the VAIC™ model (Pulic 1998, 2000, 2004, 2008). This research aims to analyse the relationship between ROE and the components of the VAIC™ model for institutions in the banking and financial sectors in Panama listed on LATINEX between 2014 and 2020. The research has a correlational scope since it aims at "understanding the relationship or degree of association that exists between two or more concepts, categories, or variables in a particular sample or context" (Hernández Sampieri et al. 2014, p. 93).

Data for constructing the VAIC™ model was obtained from Balance Sheets, Income Statements, and Notes to the Financial Statements available on the websites of the companies analysed. These data were used to calculate the value-added coefficient and the coefficients related to IC and its components: HC, SC, and RC. This process enabled the calculation of VAIC™ for each company. The procedure for calculating VAIC™ is as follows:

Table 3 details the sources from which the necessary variables for constructing the VAIC™ model were collected.

A regression model was constructed to examine the relationship between ROE, the dependent variable of this research, and the components of the VAIC: CEE, HCE, and SCE (explanatory variables). The study also includes control variables such as indebtedness (total debt/total assets ratio) and company size (natural logarithm of total assets).

The linear regression model used is given by the equation:

$$\mathrm{Y}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{X}}_1+{\mathsf{\beta }}_2{\mathrm{X}}_2+\dots +{\mathsf{\beta }}_{\mathrm{K}}{\mathrm{X}}_{\mathrm{K}}+\in$$

The coefficients ${\mathit{\beta }}_0,{\mathit{\beta }}_1,\dots {\mathit{\beta }}_{\mathit{k}}$ denote the magnitude of the effect that the (independent) explanatory variables have on the dependent Y. The coefficient ${\mathit{\beta }}_0$ represents the intercept or constant term of the model and is a constant term of the model; $\in$ denotes the error or residual term of the model.

Table 5. Hypotheses, variables, formulas, and units used.

Hypothesis	Variables	Formulas	Units
Return on Equity	ROE	(Net Income/Net Worth) *100	Percentage
VAICTM	Value Added Intellectual Coefficient	ICE + CEE	Percentage
ICE	Intellectual Capital Efficiency Coefficient	SCE + HCE	Percentage
SCE	Structural Capital Efficiency Coefficient	SC/VA	Percentage
CEE	Capital Employed Efficiency Coefficient	VA/EC	Percentage
SIZE	Total assets of the company	Logarithm of total assets	Logarithm of total assets
INDEBTEDNESS	END	(Total assets/Total liabilities) *100	Percentage

Source: Own elaboration.

Table 5. Hypotheses, variables, formulas, and units used.

Hypothesis	Variables	Formulas	Units
Return on Equity	ROE	(Net Income/Net Worth) *100	Percentage
VAICTM	Value Added Intellectual Coefficient	ICE + CEE	Percentage
ICE	Intellectual Capital Efficiency Coefficient	SCE + HCE	Percentage
SCE	Structural Capital Efficiency Coefficient	SC/VA	Percentage
CEE	Capital Employed Efficiency Coefficient	VA/EC	Percentage
SIZE	Total assets of the company	Logarithm of total assets	Logarithm of total assets
INDEBTEDNESS	END	(Total assets/Total liabilities) *100	Percentage

Source: Own elaboration.

4. Results

The descriptive statistics of the variables studied are presented in Table 6. The mean values obtained are as follows: ROE = 8.54, HCE = 5.14, SCE = 3.78, CEE = 0.08, VAIC = 9.00, INDEBTEDNESS = 0.91, and SIZE Ln_Assets = 19.82. The highest mean values were observed for SIZE, VAIC, and ROE. In terms of the standard deviation, the results for the variables were: ROE = 7.71, HCE = 2.72, SCE = 36.49, CEE = 0.07, VAIC = 36.16, INDEBTEDNESS = 0.69, and SIZE = 2.51. The variables that deviated most from the mean were SCE, VAIC, and ROE, while CEE, INDEBTEDNESS, and HCE were very close to the mean.

To determine the validity of the model, the assumptions of linearity, independence of errors, homoscedasticity, normality and non-collinearity were tested.

4.1. Linearity

Table 7 shows that there is a significant linear relationship between ROE and the explanatory variables SIZE (Ln_Assets), VAIC, and SCE (p < 0.05). The variables CEE, HCE, and INDEBTEDNESS did not show a significant linear relationship (p > 0.05) and were, therefore, not included in the model. The results indicate a significant linear correlation between the dependent variable (ROE) and the explanatory variables SCE and VAIC, as well as between CEE and INDEBTEDNESS and the logarithm of the SIZE of the assets.

4.2. Independence from Errors

The assumption of independence of errors means that the measurement errors of the explanatory variables are uncorrelated with each other. This is verified if the Durbin-Watson (D-W) statistic falls between 1.5 and 2.5. For the model of this research, the Durbin-Watson statistic was 2.065, indicating that the errors are independent. These results are shown in Table 8.

4.3. Homoscedasticity

Homoscedasticity implies that the errors have a constant variance. To verify this, a scatter plot was used with standardised predicted values (ZPRED) and standardised residuals (ZRESID). The scatter plot shows no clear pattern of association, linear or otherwise, indicating that the assumption of homoscedasticity is met, as illustrated in Figure 1.

To confirm this result, an analysis was conducted to calculate the correlation between the absolute values of the residuals and the predicted values. The results indicate no significant relationship between the residuals and the predicted values (p > 0.05), as shown in Table 9.

4.4. Normality

Normality implies that the variables follow a normal distribution. The Kolmogorov-Smirnov (K-S) test was used to assess normality, as shown in Table 10. The results indicate that the hypothesis that residuals have a normal distribution cannot be rejected (p = 0.200).

4.5. Non-Collinearity

This assumption implies that there should not be any strong linear dependence (high correlation) between the explanatory variables. To verify this assumption, the tolerance indicators and the Variance Inflation Factor (VIF) were used. In order for there to be no multicollinearity, tolerance values must be high (greater than or equal to 0.10). The VIF is the reciprocal of tolerance since lower VIF values indicate lower multicollinearity. VIF values greater than suggest serious collinearity problems. The results indicate that the assumption of non-collinearity is verified (Tolerance > 0.10 and VIF < 10 in all cases), as shown in Table 11.

4.6. Model Goodness of Fit

According to the model summary, the adjusted R² is equivalent to 0.570 (Model 2), indicating that the SCE and SIZE variables (Ln_Assets) explain 57% of the variance in ROE, as shown in Table 12. On the other hand, the F- statistic value of 99.156 indicates that this linear relationship is significant (p < 0.05), according to the results of Table 13.

4.7. Multiple Linear Regression Model

This section presents the results of the estimated model. Based on the results shown in Table 14, the regression equation in raw scores is as follows:

$$\widehat{\mathit{R}\mathit{O}\mathit{E}} \mathbf{=} \mathbf{-}\mathbf{1.34}\mathbf{+}\mathbf{0.53}\mathbf{*}\mathbf{SIZE}\mathbf{ }\mathbf{LnASSETS}\mathbf{-}\mathbf{0.15}\mathbf{*}\mathbf{SCE}$$

4.8. Goodness of Prediction

To verify the goodness of the estimate, a Z-test was applied for the means of two samples, which were considered significant for p < 0.05 values. The mean values obtained for the ROE and its estimate were 8.54 and 8.60, respectively, as shown in Table 15. Additionally, there is a significant linear correlation (p < 0.05) between the means of these variables according to the values observed in Table 16. Finally, it can be observed that there are no statistically significant differences between the mean values of the ROE and its estimate (p = 0.94), as can be seen in Table 17.

5. Discussion

The results show, according to the non-standardised coefficients (Model 2) presented in Table 13, that if the Ln_Assets SIZE remains constant, an increase of one dollar in the SCE would cause a decrease of $0.15 in the ROE. Conversely, if the SCE remains constant, a 1% increase in the SIZE Ln_Assets would result in an approximate 1% (0.53/100%) increase in the ROE. These findings align with those of Meles et al. (2016).

The standardised coefficients (β) in Model 2 indicate that the variables providing significant information (p < 0.05) for explaining ROE, in order of their weight, are (from highest to lowest) β_est.SCE = -0.72 and β_{est.SIZE Ln Assets} = 0.17). This means that the SIZE Ln_Assets variable has a positive trend, and the SCE has a negative trend (Table 14). The positive trend suggests that a 1% increase in Asset Size increases ROE by approximately 1% (0.17/100≈0.01), regardless of the SCE. The results are consistent with the findings of Meles et al. (2016) and support Barney’s (1991) resource-based theory, as larger companies managed to maintain competitive advantages through greater profitability.

The negative trend, on the other hand, indicates that for every dollar invested in the SCE, the ROE decreases by an average of $0.72, regardless of the Asset Size. In other words, investment in structural capital does not contribute to the profitability of the analysed companies. Studies conducted by Al-Musali and Ismail (2014) in the United States found no relationship between ROE and SCE. In this study, a negative relationship was found between the VAIC variable and ROE, indicating that for banks and financial institutions listed on LATINEX, higher IC is associated with lower profitability. On the other hand, studies by Al-Musali and Ismail (2014), Arslan and Kızıl (2019), and Meles et al. (2016)found a positive relationship.

Based on the results of this research, the alternative hypothesis of this study is confirmed since the SCE, SIZE, and VAIC variables are related to the ROE of the banks and financial institutions listed on LATINEX between 2014 and 2020.

6. Conclusions, Limitations, and Future Perspectives

This article analyses the relationship between the components of IC and the profitability of Panamanian banking, and financial institutions listed on LATINEX between 2014 and 2020. Utilising the VAIC^TM model, this study examines how IC influences the financial performance of these companies. By including 22 financial institutions, this research makes significant contributions, enabling managers to identify variables that add value and competitive advantages to their organisations.

The alternative hypothesis of this study was confirmed since the SCE and SIZE variables explained 57% of the behaviour of ROE in the analysed companies. The results suggest that the IC of financial sector companies listed on LATINEX is mainly affected by the SCE, which has a negative relationship with ROE, indicating that investment in structural capital does not contribute to profitability. On the other hand, larger companies achieved higher profitability during the study period. These findings significantly contribute to the scientific community by reinforcing the assumptions of the resource-based theory proposed by Barney (1991).

This study is limited to a sample of companies in the banking and financial sectors listed on LATINEX. For future research, other sectors of companies listed on LATINEX could be considered. Additionally, comparisons could also be made among different sectors on LATINEX and companies listed on other stock exchanges globally, thus providing more robust conclusions. Extending the study period to analyse the effects before and after the COVID-19 pandemic would also be beneficial. It is suggested that future studies employ other IC measurement methods, such as Skandia, M-VAIC, and Intellect, and compare the results with different methods. In addition, other profitability metrics, such as market value and Tobin's Q, could also be considered. Finally, this study serves as an important reference point for the development of future research.