Table 4.

Estimated coefficients of linear panel models.

	Fixed effects model (1)	Random effects model (2)		Pooling model (3)
	Estimate	Estimate		Estimate
Intercept	-	-16.10361	∗∗∗	-19.69592	∗∗∗
Log(GDP)	0.00666	0.36724		0.61835	∗
Log(WAGE)	1.41820	1.38012	∗∗∗	1.45152	∗∗∗
Log(UNEM)	-0.19705	-0.01797		0.14204
Log(EDU)	-0.00456	-0.25301		-0.47277	∗
R-Squared	0.35220	0.42397		0.49238

Note: ∗∗∗, ∗∗, ∗ denote significance levels on 1, 5, and 10 percent, respectively. GDP represents gross domestic product per capita, WAGE is average wage, UNEM denotes unemployment rate, and EDU is percentage of tertiary educated employees. According to the poolability test for individual cross-sectional effects using fixed effects model, all coefficients, excluding intercepts, are equal for individual effects (F = 1.432) as well as for time effects (F = 0.770). F test confirmed an existence of individual effects (F = 2.916∗∗∗) and absence of time effects (F = 1,113). There is not present cross-sectional dependence according to the Pesaran CD test (Z = -0.351). Breusch-Godfrey test for serial correlation in panel models did not confirm an existence of the serial correlation (Chi-squared=21.769). Hausman test prefers application of the (Chi-squared=3.056) fixed effects model (1).