Page 44 - ISC PROCEEDINGS 21.4

Meanwhile, Anggraeni, W (2001) argues that LF participation rates do not correlate with
EG.
2.3. The relationship between innovation and EG
Innovation is the process of creating and applying new solutions, technologies,
techniques, or management methods to improve the productivity, quality, and added
value of products, services, and processes. Several empirical studies have used the patent
application index to assess the impact of innovation on EG. Using the VAR estimator of
the GMM estimation method, Gyedu, S et al. (2021) have shown that research and
development, patents, and trademarks have a significant impact on the GDP per capita of
the G7 and BRICS countries. Pham, T. M. (2023) shows that the number of patent
applications (non-residents), R&D expenditures, and R&D staff have a positive effect on
GDP per capita.
Meanwhile, some empirical researchs show that innovation has the opposite impact
or no impact on EG. Research by Crosby, M (2000) indicates that there is an inverse
relationship between EG and the number of patents in the short term. Because patents
involve different fees that make patenting expensive in the short term. Law, S.H et al.
(2020) studying this relationship in Malaysia shows that innovation as measured by the
total number of patent applications is statistically insignificant. Pham, T.C. (2025) showed
similar results when studying in Vietnam in the period 1990-2020 using the self-regression
delay delivery model. The results of the study show that innovation reflected in the total
number of patent applications (domestic and foreign) is not yet a driving force for EG in
Vietnam
In summary, existing studies on the impact of FDI, LF, and innovation on EG show
different conclusions about the impact of these factors on EG. Therefore, this issue needs
to be studied, analyzed and evaluated cautiously.
3. Methodology
3.1. Research methodology
This study takes the following steps in turn:
Step 1: Descriptive statistics
Step 2: Unit root testing
Apply the Phillips–Perron Test (PP) to determine the order of integration for each
time series.
Step 3: Cointegration Test
Use the Johansen trace and maximum eigenvalue tests to verify the existence of
long-run equilibrium relationships among variables.
Step 4: Selection of Optimal Lag Length
Determine the optimal lag length for the VECM model using Akaike Information
Criterion (AIC) and other selection metrics
Step 5: Estimation of the VECM Model
The VECM model is used to examine the short run impact of independent variables
on dependence. The VECM model is set up as follows:
∆ = Π −1 + ∆ −1 + · · · + −1 ∆ − +1 +
1

In which: ∆ is a vector consisting of n different variables.

Step 6: Diagnostic Testing
The study carried out a number of diagnostic tests such as testing the stability of
the model and autocorrelation, heteroskedasticity in the regression model
Step 7: Impulse Response Functions and Variance Decomposition Analysis

43