Microsoft Word Thesis Gent (1). doc


  Table 2. Granger Causality Tests


Download 1.76 Mb.
Pdf ko'rish
bet29/35
Sana23.04.2023
Hajmi1.76 Mb.
#1388642
1   ...   25   26   27   28   29   30   31   32   ...   35
Bog'liq
DPTX 2010 2 0 330292 0 110731

 


53 
Table 2. Granger Causality Tests 
Pairwise Granger Causality Tests 
Date: 05/08/12 Time: 22:55 
Sample: 2002Q1 2011Q4 
Lags: 4 
Null Hypothesis: 
Obs 
F-Statistic 
Prob.
IR_SA does not Granger Cause NPL ratio 
36 
0.84247 
0.5105 
NPL ratio does not Granger Cause IR_SA 
0.97006 
0.4400 
L_CPI_SA does not Granger Cause NPL ratio 
36 
2.51689 
0.0647 
NPL ratio does not Granger Cause L_CPI_SA 
1.12151 
0.3670 
L_ER_SA does not Granger Cause NPL ratio 
36 
1.97476 
0.1269 
NPL ratio does not Granger Cause L_ER_SA 
1.45798 
0.2425 
L_GDP does not Granger Cause NPL ratio 
36 
1.90003 
0.1394 
NPL ratio does not Granger Cause L_GDP 
0.43495 
0.7822 
UNEMP does not Granger Cause NPL ratio 
36 
1.28249 
0.3014 
NPL ratio does not Granger Cause UNEMP 
0.77143 
0.5533 
From doing the pairwise Granger causality test, it is shown that granger causality runs 
from 6 variables at four lags. These results show that most of the variables are not 
significant enough to explain the changes in performance of credit risk (NPL ratio). 
Even though the theory states that it is enough to use three or four observations of 
macroeconomic variables changes, the results from the model that used four 
observations based on quarter could not observe the changes, except for inflation.
Variance decomposition 
Variance decomposition determines the contribution of each exogenous variable and 
forecasts error variance for each macroeconomic variable that explains the specific 
effects of a shock in the future. Variance decomposition shows the effect of 
macroeconomic variables (long-term interest rate, exchange rate, GDP, inflation and 
unemployment rate) to the NPL ratio. The results show that the impact of 
macroeconomic shocks on NPL ratio, start from the second quarter until the eight 
quarter (the last quarter of forecasting). Furthermore, the results show that the NPL 
ratio will be caused mostly by exchange rate in initially, (L_ER_SA) with an average of 


54 
12.75 percent (including quarters from second to sixth, see table 3). In last two quarters 
of forecasting (seventh and eighth quarters), NPL ratio gradually will be caused by 
long-term interest rate (IR_SA) with an average of 15.61 percent (including both last 
quarters, see table 3).
The forecast of variance decomposition of NPL ratio has been observed for next eight 
quarters.

Download 1.76 Mb.

Do'stlaringiz bilan baham:
1   ...   25   26   27   28   29   30   31   32   ...   35




Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling