Page 114 - Azerbaijan State University of Economics

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.81, # 2, 2024, pp. 104-116

To choose the ARDL model, the Schwarz Information Criterion (SIC) was utilised.
There was no long-term correlation between employment and human capital,
according to the particular test. Using the Ordinary Least Squares (OLS) approach,
the short-term model was estimated using the SIC selection criteria and the related
error corrective term test. The long-, short-, and short Run models' results are shown
in Table 6. Government spending on education was substantial in the near term, and
its long-term convergence adjustment speed was -0.375173, meaning that short-term
to long-term differences would be corrected at a pace of 37.51%.

In contrast to government spending on education, the long-term estimate findings
showed that government spending on health (HEXP) and life expectancy at birth
(LEB) both had negative coefficients. There is a positive long-term association
between the total labor force (LPR), inflation rate (INF), and education level (EDU).
The long-term effects of China's government spending on education are responsible
for this phenomena, since it raises the employment rate in the Chinese labor market
by polarizing the competences of university and institute graduates. This, in turn,
makes it easier for the macroeconomy to generate additional value through increased
productivity, which in turn boosts domestic demand for goods. A strong employment
participation rate enables graduates to join the Chinese labor market and contributes
to a higher employment rate, although consumption includes a high rate of inflation.
As a result, diagnostic testing were run to evaluate the quality of the model. The
outcomes of the Breusch-Pagan-Godfrey and Breusch-Godfrey Serial Correlation LM
tests are shown in Table 5.
Table.8 Diagnostic Checks

Heteroskedasticity Test: Breusch –Godfrey Serial Corr.LM
Breusch-Pegan-Godfrey Test
Prob.Value Prob.Value
F-Statistics 0.408909 0.9038 0.501091 0.6129
*
Obs R- 3.984609 0.8585 1.457574 0.4825
squared

Both the short- and long-term variants of the null hypothesis of homoskedasticity are
declined founded on the aforementioned data. This suggests heteroskedasticity in the
variables. In a similar vein, the variables' serial correlation is revealed by LM serial
correlation. According to this, the model is a decent one. The Wald Test was used to
correctly identify the combined coefficient importance of the proxies so as to
comprehend impact of the human capital proxies on employment.

114