Table 5.

Regression dynamic model (1) by using a two-step system GMM and LSDVC estimations.

Independent variables	Two-step system GMM estimation NPL is the dependent variable				LSDVC estimation NPL is the dependent variable
	(1)				(2)
	Coef.	Std Err.	Coef.	Std Err.	Coef.	Std Err.	Coef.	Std Err.
lag.NPL	0.4893***	0.0360	0.4618***	0.0299	0.0655***	0.0001	0.0650***	0.0001
LC1	−0.0107*	0.0060			−0.0507**	0.0201
LC2			−0.0116**	0.0057			−0.0241*	0.0130
BFD	0.0066	0.0080	0.0073	0.0053	0.0762***	0.0130	0.0944***	0.0122
SIZE	0.0016	0.0014	0.0010	0.0018	0.0760***	0.0238	0.0831***	0.0164
ROA	−0.4665***	0.0650	−0.4070***	0.1050	1.3302***	0.3640	1.1927***	0.2797
CAP	0.0683***	0.0147	0.0557**	0.0237	−0.1560***	0.0518	−0.1539***	0.0499
LG	−0.0015	0.0035	0.0009	0.0027	−0.0074*	0.0041	−0.0067	0.0048
INF	0.0318***	0.0080	0.0262***	0.0077	0.1771***	0.0218	0.1677***	0.0251
GDP	−0.0566***	0.0185	−0.0603***	0.0187	−0.1491	0.7325	−0.1054	0.4781
Constant	−0.0112	0.0228	−0.0019	0.0283
Observations	337		337		337		337
Instruments	59		59
Number of groups	33		33		33		33
AR (1) test	0.007		0.006
AR (2) test	0.440		0.492
Sargan test	0.287		0.275

Note: Table 5 summarizes the regression results with the NPL dependent variable by using a two-step system GMM and LSDVC estimations. Column (1) shows the GMM regressions and Column (2) reports estimated results with the LSDVC method. AR (1) and AR (2) tests show tests for correlations at lag (1) and lag (2). The Sargan test confirms the validity of instrumental variables. The LSDVC in this section, initialized by AB, is bootstrapped by 50 iterations for the standard errors (Bun & Kiviet, 2003 and Bruno, 2005). All variable definitions are displayed in Table 1. ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively.