Available online at https://journal.
com/index.
php/ijqrm/index International Journal of Quantitative Research and Modeling e-ISSN 2721-477X p-ISSN 2722-5046 Vol.
No.
3, pp.
292-297, 2025
The Effect of Macroeconomic Variables on Indonesia's Import Value Using the OLS Method Trisha Magdalena Adelheid Januaviani1*.
Kalfin2.
Aned Miranda Hutabarat3.
Nikita4.
Selvy Musdaifah5.
Nasria Nacong6 Data Science Study Program Tadulako University.
Indonesia Department of Mathematics.
Universitas Negeri Makassar.
Indonesia Department of Regional Development Management.
Bogor Agricultural University.
Indonesia Mathematics Study Program Tadulako University.
Palu.
Indonesia Corresponding author email: trishadelheid@gmail.
Abstract This study analyzes the factors influencing IndonesiaAos import value during the period 2021Ae2025 using the Ordinary Least Squares (OLS) method.
To ensure the validity of the model, a series of classical assumption tests was conducted in accordance with the Best Linear Unbiased Estimator (BLUE) criteria, including tests for normality, multicollinearity, heteroscedasticity, and autocorrelation.
The data were obtained from official publications of the Central Statistics Agency (BPS) and other relevant sources.
The estimation results demonstrate that the independent variables, namely the exchange rate (XCA), national income (XCC), foreign exchange reserves (XCE), inflation rate (XCE), and interest rate (XCI), exert varying effects on IndonesiaAos import value, with certain variables exhibiting significant influence while others remain insignificant.
The model is free from violations of the classical assumptions, thereby meeting the criteria of the Best Linear Unbiased Estimator (BLUE).
Keywords: Import Value.
OLS.
Classical Assumption Tests.
Macroeconomics Introduction The economy of a country is significantly influenced by imports, as high imports may reflect domestic demand for goods or services that cannot be supplied locally.
However, unrestrained imports can have adverse effects on the trade balance and economic stability (Ektiarnanti et al.
, 2.
International trade is one of the main drivers of a country's economy, where export and import activities play an important role in maintaining the availability of goods, increasing productivity, and strengthening economic relations between countries.
For Indonesia, imports are a vital instrument for meeting domestic needs that cannot yet be optimally produced domestically, whether in the form of raw materials, capital goods, or consumer goods (Suwardi et al.
, 2.
Over the past few decades, there has been a notable rise in imports to Indonesia.
Data from the Statistics Indonesia (BPS) indicate that IndonesiaAos import value has experienced fluctuations influenced by both internal and external factors, such as economic growth, the rupiah exchange rate, national income, foreign exchange reserves, inflation rate, and trade policies (Slamet and Hidayah, 2.
Previous research has shown a correlation between macroeconomic variables and Indonesia's import value by employing macroeconomic variables.
Hidayat et al.
found that foreign investment and government expenditure have a significant effect on imports, while the depreciation of the rupiah exchange rate does not always lead to an increase in imports.
Loso and Damanik .
demonstrated that exchange rate movements have a significant impact on IndonesiaAos imports, indicating a high sensitivity to currency fluctuations.
In general, prior research confirms that macroeconomic variables affect import values.
Therefore, this research focuses on variables that are assumed to influence import values, namely exchange rate .
cUCA), national income .
cUCC), foreign exchange reserves .
cUCE), inflation .
cUCE), and interest rate .
cUCI).
By applying multiple linear regression analysis, this study is expected to provide a comprehensive understanding of the relationships among these variables using IndonesiaAos import value data for the period 2021Ae2025.
Januaviani et al.
/ International Journal of Quantitative Research and Modeling.
Vol.
No.
3, pp.
292-297, 2025
Methods 1 Multiple Linear Regression Analysis The relationship or influence of two or more independent variables on one dependent variable is analyzed using the multiple linear regression model.
This model allows for the simultaneous evaluation of the contribution of each independent variable in explaining the dependent variable.
The form of the multiple linear regression model can be written as (Januaviani et al.
, 2.
ycycn = yu0 yu1 ycuycn1 yu2 ycuycn2 .
yuyc ycuycnyc .
yuycy ycuycnycy yuAycn ycycn ycuycnyc yu0 , yu1 , yu2 , .
, yuycy yuAycn : Dependent variable -ycn, for ycn = 1.
2, .
, ycu : Independent variable ycnyc, for yc = 1.
2, .
, ycy : Regression coefficient : Error This model remains relevant and widely used in various fields of research, including economics, social sciences, and health, due to its ability to capture complex relationships between variables.
The multiple linear regression also plays an important role in data-driven decision-making, especially when it involves accurate model parameter prediction and estimation (Montgomery et al.
, 2.
2 Ordinary Least Squares The Ordinary Least Squares (OLS) method is a commonly used technique for obtaining coefficient estimators in linear regression models.
In multiple linear regression.
OLS is used to estimate the parameter yu such that the sum of the squares of the differences between the observed values and the predicted values is minimized.
In general, the estimated multiple linear regression model can be expressed as (Burton, 2.
ycCycn = yuC0 yuC1 yuC2 .
yuCyc .
yuCycy .
with the parameter estimator yu in OLS, namely:
yuCycCyaycI = .
cU yc ycU)Oe1 ycU yc yc This method remains one of the main approaches in parameter estimation due to its simplicity, efficiency, and unbiased estimates under classical linear regression assumptions.
Recent research shows that although this method has been widely used, developments such as regularized least squares can improve estimation performance on data with high multicollinearity or a large number of predictors (Ghojogh and Crowley, 2.
3 Coefficient of Determination The coefficient of determination .
cI .
basically measures the extent to which the regression model can explain the variation in the dependent variable (Ghozali, 2.
The coefficient of determination is often used as the main indicator of model goodness.
The value of .
cI .
ranges from 0 to 1, where a value close to 1 indicates a better ability of the model to explain data variability.
The formula for the coefficient of determination is expressed as (Gao, 2.
ycIycIycI ycI 2 = ycIycIycN ycIycIycI (Sum of Squared Residual.
ycIycIycN (Sum of Squared Tota.
: yuC yc ycU yc yc Oe ycuycE 2 : yc yc yc Oe ycuycE 2 4 Classical Assumption Test In linear regression analysis, classical assumption tests are performed to ensure that the model meets the statistical requirements that guarantee Best Linear Unbiased Estimator or BLUE parameter estimation (Hansen, 2.
These tests include normality test, multicollinearity test, heteroscedasticity test, and autocorrelation test.
1 Normality Aims to check whether the residuals are normally distributed.
One commonly used method is the JarqueAeBera (JB) test (Glinskiy rt al.
, 2.
ycu ycu : sample size ycI : skewness .
a Oe .
2 yayaA = .
) y .
cIA ( .
Januaviani et al.
/ International Journal of Quantitative Research and Modeling.
Vol.
No.
3, pp.
292-297, 2025
ya : kurtosis If ycy Oe ycycaycoycyce > 0.
1, then the residuals are considered to be normally distributed (Gujarati and Porter, 2.
2 Multicollinearity This test aims to detect high correlations between independent variables.
One approach is to calculate the Variance Inflation Factor (VIF) (Ahmad et al.
, 2.
ycOyayaA = .
Oe ycIA2 ) .
ycIAA is the coefficient of determination of the regression of the ycn independent variable against the other independent If ycOyaya > 5, then multicollinearity is indicated (Ghozali, 2.
3 Heteroscedasticity This test aims to check the equality of residual variances.
One method is the BreuschAePagan (BP) test:
yaAycE = ycu y ycIA ycIA is obtained from the residual square regression against the independent variable.
If the ycy Oe ycycaycoycyce < 0.
05, there is an indication of heteroscedasticity (Wooldridge, 2.
4 Autocorrelation The test aims to detect serial correlation in residuals.
A commonly used method is DurbinAeWatson (DW) (Kim, yaycO = yu.
ceycn Oe yceycn Oe1 ) yu.
ceycn2 ) .
The value of DW close to 2 indicates no autocorrelation (Gujarati and Porter, 2.
Results and Discussion 1 Research Object The problem analyzed in this journal is to analyze the factors that influence import values in Indonesia .
cU) with a focus on economic variables, namely the exchange rate of the rupiah against the US dollar (USD) ycU1 , national income .
cU2 ), foreign exchange reserves .
cU3 ), inflation rate .
cU4 ), and interest rate .
cU5 ).
Based on the results of multiple linear regression analysis covering all independent variables, namely the exchange rate .
cU1 ), national income .
cU2 ), foreign exchange reserves .
cU3 ), inflation rate .
cU4 ), and interest rate .
cU5 ), the coefficient of determination (RSquar.
was obtained at 0.
363 or 36.
This indicates that the five variables together can explain 3% of the variation in import values in Indonesia from 2021 to 2025.
In other words, 36.
3% of the variation in import values in the analyzed data can be explained by the macroeconomic factors included in the model.
2 Classical Assumption Test 1 Normality Test Normality testing was conducted to determine whether the regression model residuals were normally This test is important because one of the classical assumptions of linear regression is that the residuals must be normally distributed (Gujarati and Porter, 2.
In this study, normality testing was conducted using graphical testing methods through Normal Q-Q Plots and Shapiro-Wilk statistical test.
Figure 1.
Normality Test Januaviani et al.
/ International Journal of Quantitative Research and Modeling.
Vol.
No.
3, pp.
292-297, 2025
Based on the Normal Q-Q Plot in Figure 3.
1, the residual points tend to follow the red diagonal line, indicating that the residual data has a distribution close to normal.
Some minor deviations are visible at the ends .
, but overall the pattern of points follows the reference line, so the assumption of normality can be accepted.
Table 1: p-values of Independent Variables Variable yec Oe yeyeCyesyenyeI ycU1 ycU2 ycU3 ycU4 ycU5 Based on Table 1, the variables that significantly affect import values at a significance level of 1% are national income .
cU2 ) .
cy Oe ycycaycoycyce = 0.
, suggesting that increases in income enhance the capacity to purchase imported Inflation .
cU 4 ) demonstrates a marginally significant positive influence at the 10% level .
cy Oe ycycaycoycyce = 083.
, implying that higher inflation is associated with greater import demand.
Conversely, the interest rate .
cU5 ) exhibits a marginally significant negative effect at the 10% level .
cy Oe ycycaycoycyce = 0.
, indicating that higher borrowing costs suppress import activities.
By contrast, the exchange rate .
cU1 ) with a ycy Oe ycycaycoycyce of 0.
38%) and foreign exchange reserves .
cUCE) with a ycy Oe ycycaycoycyce of 0.
83%) are not statistically significant, suggesting that their variations do not meaningfully account for changes in import values during the period.
2 Multicollinearity Test Multicollinearity is used to detect strong correlations between independent variables in a regression model.
One indicator used is the Variance Inflation Factor (VIF) value.
A regression model is said to be free from multicollinearity if the VIF < 5.
Table 2: Variance Inflation Factor Value ycU2 ycU4 ycU5 Based on the results of calculations using the RStudio program, the following VIF values were obtained for the independent variables: ycU2 (National Incom.
= 2.
ycU4 (Inflation Rat.
= 1.
dan ycU5 (Interest Rat.
= VIF of all values obtained are less than 5, so it can be concluded that there is no evidence of multicollinearity in the regression model of Indonesia's import value.
Thus, the independent variables in this study can be used without causing distortion in the estimation of regression coefficients (Gujarati and Porter , 2.
3 Heteroscedasticity Test The heteroscedasticity test was conducted to determine whether there was variance inequality in the residuals at each level of the independent variable in the regression model.
The heteroscedasticity test using the Breusch-Pagan Test yielded a BP value of 5.
336 with degrees of freedom .
= 3 and ycy Oe ycycaycoycyce = 0.
Since the ycy Oe ycycaycoycyce is greater than the significance level of 5% .
, it can be concluded that the regression model does not exhibit This means that the residual variance is homogeneous .
, so the regression model is suitable for further analysis.
4 Autocorrelation Test Autocorrelation is one of the important assumptions in classical regression analysis that needs to be tested to ensure the validity of the model.
Autocorrelation occurs when there is a correlation between the residuals in one period and the residuals in the previous period, which usually appears in time series data.
The presence of autocorrelation can cause the error variance to become non-constant and the estimation parameters to become inefficient, although they remain unbiased.
The results of the autocorrelation test using the Durbin-Watson (DW) test show a DW value of 1.
with ycy Oe ycycaycoycyce of 0.
The value of DW close to 2 and ycy Oe ycycaycoycyce > 0.
05 indicate that the regression model does not exhibit autocorrelation, either positive or negative.
Thus, it can be concluded that the classical assumption of no autocorrelation in the regression model is satisfied, making the model suitable for further analysis.
This finding aligns with the opinion of Gujarati and Porter .
, who state that a DW value close to 2 and ycy Oe ycycaycoycyce above the 5% significance level indicate no autocorrelation in the data.
These results are also supported by recent literature such as Januaviani et al.
/ International Journal of Quantitative Research and Modeling.
Vol.
No.
3, pp.
292-297, 2025
Wooldridge .
, which emphasizes the importance of autocorrelation tests to ensure the reliability of regression models in economic research.
5 Model of Ordinary Least Squares The regression model estimated using the OLS (Ordinary Least Square.
method produced a constant value of Ae5344.
90, a coefficient for variable ycU2 is 0.
0087, variable of ycU4 is 1046.
63, and variable of ycU5 is Ae556.
Before the model was used for interpretation, a series of classical assumption tests were conducted, including normality, multicollinearity, heteroscedasticity, and autocorrelation tests.
The test results showed that all assumptions were met:
the residual data were normally distributed, there were no signs of multicollinearity between independent variables, no heteroscedasticity was found, and there was no autocorrelation.
Thus, the regression model obtained meets the BLUE (Best Linear Unbiased Estimato.
criteria, so that parameter estimates can be considered valid and reliable for use in further analysis.
The resulting regression model is as follows:
ycU = Oe5344.
0087ycU2 1046.
63ycU4 Oe 556.
72ycU5 yce The regression model shows that the constant value of -5344.
90 does not have practical meaning, as import values cannot be negative, but it functions as a mathematical starting point.
The coefficient for national income .
cUCC) is 0.
meaning that an increase in national income increases IndonesiaAos import values.
Inflation .
cUCE), with a coefficient of 63, also increases import values.
In contrast, the interest rate .
cUCI), with a coefficient of -556.
72, decreases import The residual .
represents other factors not included in the model that influence import values.
Overall, this model suggests that variables XCC and XCE drive an increase in import values, while ycUCI actually suppresses import values.
Conclusion Based on the description in the previous chapter, the following conclusions can be drawn:
The regression model used has met all classical assumption tests, namely normality, multicollinearity, heteroscedasticity, and autocorrelation tests.
This indicates that the model formed has met Best Linear Unbiased Estimator or BLUE criteria so that the parameter estimation results can be trusted and are valid for further analysis.
There were five macroeconomic variables tested .
xchange rate, national income, foreign exchange reserves, inflation, and interest rate.
, but only three variables were retained in the final model, namely national income .
cUCC), inflation .
cUCE), and interest rates .
cUCI).
Meanwhile, exchange rate .
cUCA) and foreign exchange reserves .
cUCE) were excluded from the model because they were found to be statistically insignificant in the normality test.
Based on the OLS model, the national income variable .
cUCC) was found to have a significant positive effect on Indonesia's import value, meaning that the higher the national income, the greater the demand for imports.
The inflation variable .
cUCE) had a positive and significant effect at the 10% level, indicating that an increase in inflation drives an increase in imports.
Conversely, the interest rate variable .
cUCI) has a negative effect on import value, indicating that an increase in interest rates can suppress import activity.
References