The External Economic Environment as an Attractive Factor for Foreign Direct Investment in Kosovo: An Econometric Analysis

FDI is a form of investment where a foreign company or individual invests directly in a business or asset located in another country, taking control or at least a significant stake in the management of the business and this investment may include purchasing shares in a foreign company, establishing subsidiaries, or purchasing real estate and other activities that require direct involvement and commitment in the management of the business [43].

In this paper, the main theories related to the influence of macroeconomic factors on FDI include several concepts and authors who have contributed to the development of these theories. Dunning's theory [2], known as the Eclectic Paradigm or OLI, mentions three key advantages that a firm must have to invest abroad: Ownership Advantage (O), Location Advantage (L), and Internalization Advantage (I). Dunning emphasizes the importance of location factors such as INF, tax policies, and banking infrastructure that play an important role in attracting FDI, thus supporting investors' demands for a stable economic and institutional environment.

Hymer's theory [1] focuses on the competitive advantages of international firms, arguing that they invest in foreign markets to avoid local competition and to exploit advantages that cannot be realized in domestic markets. He emphasizes that foreign investment is driven by the desire to create dominant positions in international markets.

This study draws on Dunning's Eclectic Paradigm (OLI framework) to examine FDI decisions, emphasizing location-specific factors such as inflation stability, taxation policies, and banking infrastructure. These factors align with the paradigm's location advantages, showcasing Kosovo's potential to attract foreign investors.

Dunning and Hymer are two important figures in the development of the multinational enterprise (MNE) theory and have contributed to the understanding of the factors that drive the growth and international expansion of these firms.

Hymer was one of the pioneers in the theory of multinational firms. He argued that firms operating in international markets possess unique advantages, such as the ability to avoid local competition and create a dominant position. He mentioned that MNEs are motivated to invest abroad in order to spread their competitive advantages and escape from the limitations and competition of local markets [1].

Hymer's theoretical framework (Table 1) emphasizes the significance of market imperfections, firm-specific advantages, and barriers to internationalization. The variables in the study align closely with these theoretical aspects, providing a clear context for understanding the motivations of foreign investors.

Table 1. Hymer's theoretical framework and its connection to the variables in your study on the motivation of foreign investors in Kosovo

Dunning [2, 3, 44, 45] introduced the Eclectic Paradigm (OLI Theory), combining prior theories to explain why MNEs invest abroad:

- Ownership Advantage (O): Competitive assets like technology, brand, or human capital.

- Location Advantage (L): Host-country benefits like low labor costs, market access, or favorable policies.

To assess the influence of independent variables on the dependent variable (MFI) and examine the relationships between variables, a series of estimation steps were followed. Initially, a descriptive statistical analysis was performed to summarize the main features of the data. Next, demographic analysis was carried out, with the study data being visualized through graphs to emphasize key demographic characteristics. Subsequently, diagnostic tests were conducted to verify the reliability and suitability of the data, ensuring that valid statistical conclusions could be drawn. Following this, Kendall's Tau-b analysis was employed to measure the relationships between variables. Finally, MLR analysis was applied to estimate the structural equation model and assess the impact of the independent variables on the dependent variable (MFI).

4.1 Descriptive statistics

Key statistical measures (Table 4), including the number of observations, mean, standard deviation, Skewness and Kurtosis, were calculated for each variable: To identify the most suitable variables for analyzing the motivation of foreign investors in Kosovo, several aspects of descriptive statistics are considered. Stability and Reliability: Variables with lower variability, such as interest rate (SD: 0.891), AC/loans (SD: 1.073), and taxes (SD: 1.058), are more stable and reliable for analysis. Response Distribution: AC/loans and interest rate have lower skewness and are closer to a normal distribution, indicating a more consistent perception among investors. Investor Perception: Variables with higher averages, such as AC/loans (6.07) and investor motivation (25.84), show a positive perception and significant impact.

Table 4. Descriptive statistics

Negative Kurtosis: Variables such as ER, INF, and E-BBS exhibit a flatter distribution, requiring further analysis. Interest rate, access to credit/loans, and taxes are the most reliable variables, while investor motivation remains a key factor.

4.2 Demographic statistics

FDI and other variables are presented in graphical and tabular form below.

Based on data from the Central Bank of Kosovo (CBK) for the period from January to September 2024 (Figure 3), FDI reached €657.3 million, marking an increase of €38.5 million or approximately 6.2% compared to the same period in 2023. However, compared to previous years, there is a decline in the annual growth trend.

3.png

Figure 3. FDI by year (millions of euros) and annual increase/decrease in %

According to data from January to September 2024 (Figure 4), FDI continue to be predominantly dominated by real estate activities, accounting for 71.4%, followed by financial and insurance activities at 14.9%. The remaining 13.7% consists of other sectors, including construction at 1.6%, manufacturing at 1.5%, mining and quarries at 1.4%, wholesale and retail trade, motor vehicle repair, and motorcycle repair at 1.3%, among others. Motivation of foreign investors in Kosovo.

4.png

Figure 4. FDIs by year and economic activities in %

According to analysis by country of origin, the leading investors in Kosovo are from the EU 27 (Figure 5). FDI in Kosovo by country (net) for the period from January to September 2024 are as follows: Germany 20.8%, Switzerland 20.0%, the USA 9.5%, Albania 11.4%, Austria 7.0%, the United Kingdom 2.3%, among others, representing their share in total FDI.

5.png

Figure 5. Origin of growth investment

6.png

Figure 6. Motive of investors and the performance of economic factor's

The opinion of foreign investors on economic factors is presented in the created Figure 6.

The results presented in Figure 6, based on the statistical analysis of the motives of foreign investors, suggest that although various economic factors such as ER, INF, IR, AC, CO, e-banking and banking services, economic growth, and taxes are related to these motives, their impact is moderate in this sample. The low mean and standard deviation indicate that the data distribution is compact, with low variability among the data values.

This may imply that the analyzed factors do not have a high impact on the motives of foreign investors. To better understand the potential impacts, it may be necessary to apply more advanced statistical methods or consider additional factors that were not included in this analysis.

7.png

Figure 7. The impact of 8 independent variables on the dependent variable MFI

Figure 7 presents the results of the linear regression analysis, indicating that economic factors influence the motivation of foreign investors to invest in Kosovo. The regression equation Y=25.5+6.8X reveals a positive relationship between the economic variables and investor motivation. The R² value of 45% suggests that 45% of the variation in investor motivation can be explained by these economic factors. The scatter plot displays the distribution of motivation based on the predicted values. To boost investments, it is recommended to improve economic conditions, particularly by enhancing AC and stabilizing ER.

4.3 Statistical tests

To create the motivation index (Dependent variable), data from the components PCA have been used, and the formulas and calculations for each of the 6 components we extracted were processed. The loadings for each variable in the components have been used and multiplied by the weights provided.

4.3.1 Testing dependent variables

Calculations, Cronbach's Alpha, KMO test and PCA for each component Dependent variable MFI (See: This part was not included in the article due to the volume of words that were added too much):

a) A Cronbach's Alpha (Table 5)

Value of 0.849 indicates that the 16 items in the foreign investor motivation index demonstrate high internal consistency and reliably measure the same underlying construct.

Table 5. Reliability statistics-MFI

b) KMO and Bartlett's Test (Table 6)

With a KMO Table 6 value of 0.745: The sample is suitable for factor analysis, because the value falls into the "good" category. However, Bartlett's Test of Sphericity is statistically significant (p<.001), which means that the correlations between items are sufficient to proceed with factor analysis.

Table 6. KMO and Bartlett's test-MFI

c) Tests PCA-MFI

Table 7. Component matrix^a

Table 7 identifies which variables contribute most to each of the 6 main components, with the aim of building a common index that measures the motivation of foreign investors in Kosovo.

Summary of Key Components

-Component 1 – Local Support and Positive Climate:

Positive: Support from Kosovar businesses, FDI from other markets, low CO.

Negative: Personal motive for investment, natural resources. Reflects the influence of the supportive local climate on motivation.

- Component 2 – Institutional Access and Local Market:

Positive: Access to local markets, economic policy, business environment.

Negative: Albanian culture, regulations for foreign businesses. Indicates the importance of institutional structures and perception of the environment.

- Component 3 – Regulatory Framework and Social Impact:

Positive: Import regulations, labor force, family impact. Reflects regulatory ease and social impacts on decision-making.

- Component 4 – Access to Information and Market Size:

Positively: Market information, market size, government regulations

Assess easy access to information and the size of opportunities.

- Component 5 – Cost and Benefit:

Positively: Size of investment, low competition.

Negatively: Labor force, market information. Measures the balance between cost and benefit in investment.

- Component 6 – Institutional Support:

Positively: Government care for exports.

Negatively: Influence from family/friends. Related to the role and trust in the government in attracting investment.

These components were then used to calculate the overall investor motivation index, based on their scores and the weight of each factor. This approach has also been applied previously [94] demonstrating the importance of PCA testing in reducing insignificant variables

The result of the main components (component scores) calculated based on the average ratings and the matrix of components:

The Component Transformation Matrix (Table 8), for MFI shows the correlations between the original factors and the newly extracted factors after applying Varimax rotation. High positive loadings (e.g., 0.860 for Component 1) indicate a strong relationship between the original variable and the new component. Negative values (e.g., -0.177 for Component 2 on Component 1) suggest an inverse correlation. The matrix helps to identify which components are strongly associated with each extracted factor, providing clearer interpretation of the data. The use of PCA and Varimax rotation makes it easier to understand the underlying dimensions of the data by maximizing variance explained per component.

Table 8. Component transformation matrix-MFI

Component 1: Low market competition; Easy access to local markets; Component 2: Healthy economic policies; Weak business environment in Kosovo; Component 3: Availability of labor force; Simple import regulations; Component 4:Ease of access to market information; Simple government regulations for foreign businesses; Component 5: Market size available; Availability of natural resources; Component 6: Government care for export promotion initiatives.

This index represents a summary of the influence of each component on motivation, considering the loadings of the respective variables and the weights we assigned.

A scree plot is used to visualize the results of a PCA and helps in determining the number of components that should be retained in the model (Figure 8). This plot shows the eigenvalues (the amount of variance explained) for each component and typically uses a "Kaiser criterion" to decide the optimal number of components. In a scree plot, components with an eigenvalue greater than 1 are considered significant and should be retained, while those with an eigenvalue less than 1 can be excluded.

8.png

Figure 8. Screen plot-the reduced variables with the most impact on investor motivation

MFI Index based on these 16 factors, we can use a simple weighted sum formula or an average-based approach, depending on how you want to weigh the factors. No, there is no need for standardization to create an index with the average of Likert scales (when the variables are on the same Likert scale). Using the mean of Likert data, where each value has the same importance across all questions, is sufficient to obtain an overall summary without the need for standardization of variables. In addition to using PCA for the independent variables, it should be noted that PCA was used (also for the dependent variable since it is represented by 16 variables to identify the main factors that measured the motivation of foreign investors in Kosovo), then an index was created based on these components, with the weights resulting from PCA. Furthermore, further validation is always necessary, internal consistency (Cronbach's Alpha) to verify that the components are reliable, which has been done.

Formula for the Index: If we assume equal weights for each factor (which is common when there are no specific reasons to assign different weights), the formula for the MFI Index would be:

This will give you an average score for all the factors, which can be interpreted as the overall MFI Index.

Table 9. Using weights from PCA

Table 9 presents the key factors influencing the motivation of foreign investors in Kosovo, where each factor is assigned a rating by respondents and a weight obtained from Principal Component Analysis (PCA). By multiplying each rating by its corresponding PCA weight and summing these weighted scores, a total value of 10.564 is obtained.

- Formula for Motivation of Foreign Investors (FDI) Index

Applying this to your factors:

$\begin{aligned} M F I_{ {index }}=\frac{1}{16} \sum_{i=0}^n & \text { (Easy access to local markets + ... } + \text { Rating of FDI from other markets }+\ldots+\text { ) }\end{aligned}$                        (22)

The Motivation of Foreign Investors Index (MFI_index) formula then divides this total by the number of factors (n=16) to calculate the average motivation score:

$\begin{aligned} & M F I_{ {index }} \\ & =\frac{1}{n} \sum_{i=0}^n\left(\frac{\binom{\text { Weight PCA1 } * \text { Rating } 1+\ldots}{+ \text { WeightPCA16 } * \text { Rating } 16}}{(\text { Weight } 1+\ldots+\text { Weight } 16)}\right)=\frac{10.564}{16}=5.11\end{aligned}$                         (23)

Final result: MFI Index=5.11 (This result indicates a relatively high level of investor motivation in relation to local support, positive business climate, and economic factors).

In summary, using PCA-derived weights assigns statistical significance to each factor rather than relying solely on subjective ratings, resulting in a more reliable and objective measure of investor motivation.

This value was transferred to SPSS 25 to the dependent variable MFI.

4.3.2 Testing independent variables

Statistical tests are essential for assessing data reliability, validity, Cronbach's Alpha evaluates internal consistency and reliability. ANOVA with Friedman’s Test is used to detect differences between related samples, and KMO and Bartlett's Test assess data suitability for factor analysis. Lastly, PCA reduces data dimensionality while retaining key information for analysis.

a) The Cronbach's alpha coefficient (CORR) test

Table 10, Reliability assessed by looking at the mean, standard deviation, and sample size. Mean: Both variables (MFI and Economic Factors) show positive evaluations, with means around 4.5 and 5, suggesting favorable opinions from respondents. Standard Deviation: The small standard deviations (0.25923 and 0.28588) indicate consistent responses, with little variation and a high level of agreement. Sample Size: A sample size of 148 is sufficient for reliable analysis and allows for generalization of the results.

Table 10. Item statistic

Instrument Reliability Analysis (Reliability Statistics) Table 11: To assess the reliability of the questionnaire used in this study, Cronbach's Alpha coefficient was applied, which measures the internal consistency of a group of items that are supposed to measure the same theoretical construct.

Table 11. Reliability instrumental statistics (Cronbach’s alpha test)

The Cronbach's Alpha value=.769 indicates a good level of reliability, as in the literature values above .70 are considered acceptable for research analysis [75]. The Alpha value based on standardized items (.787) is slightly higher, which indicates that standardization of items may have slightly increased uniformity among them. The number of items included in the analysis is 9, which is an adequate number for a simple factorial structure.

Based on these values, the instrument used to measure factors affecting foreign investment in Kosovo shows a good degree of internal consistency and is suitable for use in further statistical analyses.

b) ANOVA with Friedman's Test

Table 12. ANOVA with Friedman's test

Grand Mean=4.7710

a. Kendall's coefficient of concordance W= .478.

Table 12 presents the results of Friedman's ANOVA test.

Key points: Friedman's Chi-Square=100.568, p=0.000: There is a statistically significant difference between the items. Kendall's W=0.478: Indicates a moderate level of agreement among participants on how they rank the items.

Grand Mean=4.7710: The overall average score across all items. The results show that the items differ significantly, with moderate agreement in rankings among respondents.

c) Factor Analysis-KMO and Bartlett's Test

The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy is a test developed by Kaiser [95] in 1970. This test is used to assess the suitability of data for factor analysis. Kaiser [95] was a psychologist and statistician who contributed to the development of statistical techniques, and one of his most well-known contributions is the development of this index to evaluate the adequacy of a sample for factor analysis.

Factor Analysis Results Interpretation:

i) Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy

The KMO (Table 13) value is 0.707, which indicates a moderate level of sampling adequacy. A KMO value between 0.5 and 0.7 is considered acceptable, suggesting that the data is suitable for factor analysis, but there may still be some room for improvement.

Table 13. KMO and Bartlett's test

ii) Bartlett's Test of Sphericity

The Chi-Square (Table 13), value is 732.42 with 36 degrees of freedom (df), and the Sig. value is 0.000. This result is statistically significant (p<0.05), meaning that the correlation matrix is not an identity matrix, and there are sufficient correlations among variables to perform factor analysis.

c) PCA tests

Table 14 indicates how much of the variance in each variable is explained by the extracted factors. Variables with high communalities, such as the ER (0.849) and CO (0.801), are strongly related to the factors.

In contrast, the tax variable has a lower communality (0.551), indicating a weaker relationship with the factors. This suggests that the extracted factors adequately describe the data for most of the variables.

Table 14. Communalities

Table 15 presents the variance explained by each component in PCA, used by Kida et al. [96], originally introduced by Hotelling [97]. The first three components explain 73.09% of the total variance, suggesting that they are the most important factors in understanding the data. The remaining components (4-9) explain a smaller portion of the variance, indicating they have less influence on the overall data structure. The rotation process redistributed the variance explained by each component, but the overall variance explained by the three key components remained the same. Overall, the PCA suggests that the first three components are the most critical factors in understanding the underlying structure of the data, and the other components do not significantly contribute.

Table 15. Total variance explained

As shown in Figure 9, the first two components number Economic factors (faktoret ekonomik) represent the majority of the variance, specifically 45.9% and 15%, which together account for 60.9% of the total variance. This indicates a significant reduction in the dependency between the original variables, transforming them into independent components. This result is crucial as it captures a large portion of the information, making the analysis process more efficient and reliable. Using these independent components, which hold the majority of the variance, helps eliminate the potential impact of multicollinearity and improves the accuracy of the analytical models that will be used later.

9.png

Figure 9. Scree Plot component number (variables factor economic and MFI)

Table 16. Rotated component matrix^a

Table 16 presents the rotated component matrix from PCA [98], highlighting three key components influencing foreign investor motivation in Kosovo:

• Component 1: Traditional Macroeconomic Factors

This component includes variables that reflect the macroeconomic stability and policies that are fundamental for attracting foreign investment. The key factors are the ER (.778), interest rate (.862), CO (.775), and economic growth (.652). These elements are considered traditional macroeconomic factors, where economic stability, favorable financial rates, and a competitive environment are essential for encouraging foreign investors.

• Component 2: Access to the System and Financial Facilities

This component focuses on access to modern financial services and banking facilities, which are crucial for fostering a conducive investment environment. The factors included are INF (.840), e-banking and banking services (.803), and AC (.523). This component highlights the role of financial system development and digitalization in shaping the business climate for foreign investors.

• Component 3: Fiscal Aspects and Motivation for Business

This component emphasizes factors directly impacting the decision to start a business in Kosovo. It includes the motivation to start a business in Kosovo (-.801), taxes (.704), and economic growth (.511). It reflects the importance of both personal motivation and the fiscal environment, where the tax burden and the perception of Kosovo's economic attractiveness significantly influence foreign investors' decisions.

Table 17. Component matrix^a

Table 17 presents the component matrix, where three components are extracted to identify the key factors influencing foreign investors' motivation in Kosovo. The first component is associated with macroeconomic factors, including ER, AC, and CO. The second component highlights investor motivation, showing a negative impact from IR and slow economic growth. The third component is related to financial services and taxes, along with the negative effects of INF and IR. This analysis underscores the significant role that economic factors and financial services play in shaping foreign investors' motivation.

• Anti-Image Matrix Interpretation

All diagonal values must be above 0.5 to be considered suitable for inclusion in the factor analysis.

The Anti-Image Correlation Matrix (Table 18), was used to evaluate the adequacy of each individual variable for factor analysis. The diagonal elements represent the Measures of Sampling Adequacy (MSA) for each variable. All MSA values are above the minimum acceptable threshold of 0.5, ranging from 0.633 (Economic growth) to 0.824 (Motivation to start a business in Kosovo). This indicates that all variables contribute adequately to the factor structure and are suitable for inclusion in the analysis.

Table 18. Anti-Image correlation matrix (Diagonal MSA values)

However, some high off-diagonal correlations (e.g., between "ER level" and "CO": -0.727) suggest potential multicollinearity or overlap between variables, which should be considered during factor interpretation.

The Anti-Image Matrix, together with the KMO value of 0.683 and the statistically significant result of Bartlett's Test of Sphericity (Chi-Square =732.42, df=36, p< .001), supports the appropriateness of applying factor analysis on this dataset.

In the component plot (Figure 10), from p3 (MFI) to 8 independent variables (F1-F8), the nodes illustrate the relationship between the original variables and the components created after PCA. Each component explains a significant portion of the variance, and the nodes help identify the influence of each variable on the independent components. Variables that are closely related to each other are grouped together, while those that contribute more to the overall variance appear more distinct in the component plot. In this way, the created components help reduce the dependency between variables and highlight the key factors influencing the analysis results.

10.png

Figure 10. Component plot–Analysis weighted by economic factors

4.4 Testing study hypotheses

Hypothesis testing in this research uses multifactorial linear regression, specifically the Ordinary Least Squares (OLS) method, to measure the impact of economic factors on foreign investors' motivation in Kosovo [94]. The regression model applied is MLR, with multiple independent variables predicting the dependent variable.

4.4.1 Results of the multiple linear regression analysis for the first hypothesis

The MLR model elucidates how macroeconomic factors influence foreign investor motivation (MFI). It demonstrates that variables such as taxes, IR, INF, CO, modern banking services, economic growth, AC, and ER significantly affect MFI. Understanding these relationships is crucial for policymakers and businesses aiming to attract foreign investment and foster economic development.

The first Hypothesis H1: ER, INF, IR, AC, CO, e-banking and banking services (E-BBS), economic growth (EGr), and taxes (Tax) have a positive and statistically significant impact on the return of the motivation of foreign investors in Kosovo.

The represents the model summary for a regression analysis Table 19, specifically a multiple regression model. The MLR model indicates a moderate relationship between independent factors and the motivation of foreign investors (MFI), with an R value of 0.562. The R² value of 0.316 suggests that 31.6% of the variability in MFI is explained by the factors in the model. However, the adjusted R² of 0.277 implies that some variables may have limited impact. The F-statistic (Sig. F=0.000) confirms the overall statistical significance of the model.

However, the Durbin-Watson statistic of 0.366 indicates moderate positive autocorrelation, raising concerns about the reliability of the model for analysis. Durbin-Watson and Autocorrelation, as seen in Table 13, are more applicable to time-series data, where values are linked to different time periods for the same observational unit. In this study, the survey data are not time-series, but are collected from different individuals at the same time, and therefore, using the Durbin-Watson statistic (e.g., value 0.366) is not appropriate in this context. Instead, it is important to focus on the P-values for the coefficients in the model, which were analyzed to identify which factors truly influence the motivation of foreign investors (IMF). To support the reliability of the results and to verify the accuracy of the statistical methods used, several other important tests were also employed, which include: Cronbach’s Alpha Test: This test was used to assess the reliability of the scales used in this study. The Cronbach’s Alpha value helped measure the internal consistency of the questions and variables used. PCA was used to uncover the structure of the variables and to create a new variable for investor motivation. PCA helped reduce the dimensionality and focus on those components that explain the most variance in the data. KMO Test (Kaiser-Meyer-Olkin): This was used to check the adequacy of the data for PCA. The KMO test value helped evaluate the suitability of the sample for factor analysis. Barlett’s Test: This test was used to assess whether the data are sufficiently correlated to justify the use of PCA. A positive result from the test confirmed that the data were suitable for PCA. The use of these important tests helped ensure the quality of the model and supported the interpretation of the factors influencing the motivation of foreign investors, providing a reliable and robust analysis of the data.

Multicollinearity (VIF) (Table 19): Evaluated to determine whether some factors are highly correlated with each other, which may affect the stability of the model. To assess multicollinearity among the independent variables, Collinearity Statistics (Table 19), were examined using Tolerance and Variance Inflation Factor (VIF) values. All VIF values were found to be below the critical threshold of 4.0, ranging from 2.195 to 3.805, while Tolerance values remained above 0.2. These results indicate that multicollinearity is not a serious concern in the regression model and does not distort the estimation of coefficients. Therefore, no corrective measures such as variable elimination or transformation were deemed necessary.

Table 19. Coefficients^a

To assess the reliability of the instruments used, a reliability analysis was performed with Cronbach‘’s Alpha( Table 11), which measures the internal consistency of the survey questions. Also, Kaiser-Meyer-Olkin (KMO) (Tables 13 to Table 14) were used to assess the suitability of the data for factor analysis and PCA table to reduce the dimensions and identify the structure of the data. These analyses show that the instruments are reliable and appropriate for your study.

F Change: The value of 8.038 is the F-statistic (Table 20), which tests whether the regression model as a whole is statistically significant. Since it is relatively high, this suggests that the model fits the data well and that the independent variables jointly have a significant impact on the dependent variable. df1: This represents the number of predictors in the model, which is 8. These are the factors listed under the "Predictors" section. df2: This represents the degrees of freedom for the residuals, which is 139. It is calculated as the total number of observations (147) minus the number of predictors (8) minus 1 for the intercept.

​Table 20. Model summary^b

Overall, the table suggests that the regression model explains a significant portion of the variation in investment growth, and the predictors together are significantly related to the dependent variable. However, since R² is 0.316, there may still be other factors influencing investment growth that are not included in this model.

The ANOVA results (Table 21) show that the regression model is statistically significant (F=8.038, p=0.000), indicating that the independent variables (taxes, IR, INF, CO, e-banking, economic growth, AC, ER) collectively influence foreign investors' motivation in Kosovo. The Regression Sum of Squares (6792.157) explains the variance, while the Residual Sum of Squares (14682.843) reflects unexplained variance. The strong F-value indicates a good model fit, and the p-value confirms the statistical significance of all variables. In conclusion, these factors significantly impact foreign investment decisions in Kosovo.

Table 21. ANOVA^a

Residuals Statistics (Table 22) evaluates the regression model's quality for foreign investors' motivation in Kosovo (MFI). The predicted values range from 8.77% to 40.71%, with a mean of 25.5%, reflecting the distribution of investor motivation. Residuals range from -18.61% to 18.51%, with a mean of 0.00%, showing no systematic bias. Standardized residuals fall within the normal range (±2), indicating no extreme outliers. The small standard deviation of residuals (9.99%) suggests a balanced error distribution. These results confirm a good model fit and normal error distribution, explaining factors like INF, IR, credit access, and taxes. The data used in this study were collected through a structured questionnaire using a Likert scale, where respondents rated the importance of various factors influencing foreign investor motivation on a scale from 1 (not important at all) to 5 (very important).

Table 22. Residuals statistics^a

To assess multicollinearity among the independent variables, Collinearity Statistics (Table 19) were examined using Tolerance and Variance Inflation Factor (VIF) values. According to general econometric rules, a VIF value above 10 is considered a strong indicator of problematic multicollinearity, whereas values below 5 are generally acceptable [94]. The cross-sectional data do not make it necessary to use VIF; other tests, which are mentioned earlier in the study, are more appropriate for assessing the reliability of data from structured questionnaires. However, since the MLR model has been used and VIF is part of it, its importance is not overlooked. It can be seen that the highest VIF value does not exceed 4.512, which indicates that it does not render the analysis unreliable.

$\begin{gathered}

\mathrm{MFE}=30.363+2.584 * \mathrm{X}_{(E R)}+2.734 * \mathrm{X}_{(\mathrm{INF})}+ \\

(-3.144) * \mathrm{X}_{(\mathrm{IR})}+(-5.575) * \mathrm{X}_{(\mathrm{ACt})}+(- \\

1.210) * \mathrm{X}_{(\mathrm{CO})}+(-.643) * \mathrm{X}_{(E-\mathrm{BBS})}+6.497 * \\

\mathrm{X}_{(\mathrm{EGr})}+(-.077) * \mathrm{X}_{(\mathrm{Tax})}=30.8 \%

\end{gathered}$                           (24)

In Table 19, the MLR model explains the Motivation of Foreign Investors (MFE) based on macroeconomic variables. This model shows that some variables, such as economic growth, INF, and ER, have a positive and strong impact on attracting investors, while factors like IR, AC, CO, and taxes have negative effects. To enhance the motivation of foreign investors, these factors must be managed to create a more attractive investment environment. However, only 30.8% of the variation in their motivation depends on these factors, indicating the presence of other influencing elements not included in the analysis. Here is an explanation for each variable:

Constant (30.363): This is the intercept or base value of the dependent variable (MFE) when all the independent variables are zero. In other words, if all the economic variables (like ER, INF, etc.) are zero, the model predicts a baseline motivation of 30.363 for foreign investors. ER (2.584): For every unit increase in the ER, the foreign investors' motivation (MFE) increases by 2.584 units, assuming all other variables remain constant. This suggests that a higher ER positively impacts foreign investment motivation. INF (2.734): For every unit increase in INF, the motivation of foreign investors increases by 2.734 units, assuming all other factors remain unchanged. This shows that INF might be associated with higher motivation for foreign investment in Kosovo. IR (-3.144): For every unit increase in IR, the foreign investors' motivation decreases by 3.144 units. This negative relationship indicates that higher IR tend to deter foreign investors, likely due to increased borrowing costs. AC (Beta: -5.575): This coefficient suggests that more limited access to credit has a significant negative impact on investor motivation. The very small p-value (0.000) indicates that this factor has a strong and economically significant influence. When investors face difficulties in securing the necessary financing, they may have less motivation to invest. CO (-1.210): As competition in the market increases, foreign investors' motivation decreases by 1.210 units. High levels of competition might discourage investors, as they perceive it as harder to secure market share or achieve profitability. E-banking and Banking Services (-0.643): A rise in E-BBS or the quality of banking services is associated with a decrease in investor motivation by 0.643 units, which could imply that investors are less reliant on banking services, or that better services might lead to lower perceived risks, reducing the need for high investor motivation. Economic Growth (6.497): For every unit increase in economic growth, foreign investor motivation increases by 6.497 units. This highlights the strong positive impact of a growing economy on attracting foreign investment. Taxation (-0.077): For every unit increase in tax rates, foreign investor motivation decreases by 0.077 units. Although this effect is relatively small, it suggests that higher taxes might discourage investment. Overall Summary: The variables with the strongest and most significant positive impact on foreign investor motivation are Economic Growth (Beta=6.497), ER (Beta=2.584), and INF (Beta=2.734).

This indicates that economic factors such as economic growth, investment opportunities, and ER stability are crucial for attracting foreign investors. On the other hand, IR, AC, CO, and Taxes have a negative impact on investor motivation. The model suggests that fostering economic growth and managing inflation could be effective in boosting foreign investment, while reducing barriers such as high IR and taxes could further enhance investment attractiveness.

4.4.2 The testing second hypothesis

Hipoteza e dytë testohet përmes testit jo-parametrik Kendall's Tau_b.

To analyze the relationship between ordinal variables, this study applies the non-parametric statistical test Kendall's tau_b. This test assesses the strength and direction of the association between two ordinal variables, while also accounting for tied ranks, making it more appropriate than Spearman's rho in certain cases, applied also by Kida [99]. The coefficient value ranges from -1 to +1, where positive values indicate a direct relationship, and negative values indicate an inverse one. This test was used to evaluate the second hypothesis, which concerns the relationship between the motivation of foreign investors and the factors related to the economic environment in Kosovo.

H2: There is a positive and statistically significant relationship between ER, INF, IR, AC, CO, e-banking and banking services, economic growth, and the motivation of foreign investors in Kosovo.

Table 23. Correlation Kendall's tau_b

Based on the Kendall's Tau-b correlation table (Table 23), the following variables exhibit statistically significant correlations with Foreign Investor Motivation (MFI) at the 0.05 significance level:

i) ER: τ=0.154, p=0.026;

ii) AC: τ=0.177, p=0.012;

iii) CO: τ=0.135, p=0.044.

These findings suggest that MFI is positively associated with ER, AC, and CO. However, the correlation coefficients are relatively small, indicating weak associations. Nonetheless, the statistical significance of these correlations implies that they are unlikely to be due to random chance.

The other variables—INF, IR, E-banking & Banking Services (EBSSB), and Economic Growth Rate (EgR)—do not show statistically significant correlations with MFI, as their p-values exceed the 0.05 threshold.

In summary, while MFI is weakly correlated with ER, AC, and CO, these relationships are statistically significant, suggesting that these factors may influence foreign investor motivation. However, the weak strength of these correlations indicates that other unexamined variables may also play a significant role in determining MFI.

These findings suggest that improving AC and stabilizing the ER can positively influence foreign investor motivation. Additionally, managing interest rates and inflation can create a more favorable investment climate.

In summary, the Kendall's Tau-b correlation analysis supports the regression results, emphasizing the importance of macroeconomic stability and financial accessibility in attracting foreign investments. Based on these points, the second hypothesis can be partially supported, as most of the variables show positive and statistically significant correlations. However, some variables, such as interest rates and E-BBS, do not fully align with the hypothesis, which weakens its complete support.