APPENDIX
UNIT ROOT
EXR at Level Form
Null Hypothesis: EXR is stationary
Exogenous: Constant
Bandwidth: 3 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.691988
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 1510.165
HAC corrected variance (Bartlett kernel) 5228.928
KPSS Test Equation
Dependent Variable: EXR
Method: Least Squares
Date: 01/31/25 Time: 20:16
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 408.5417 8.103045 50.41829 0.0000
R-squared 0.000000 Mean dependent var 408.5417
Adjusted R-squared 0.000000 S.D. dependent var 39.69665
S.E. of regression 39.69665 Akaike info criterion 10.24118
Sum squared resid 36243.96 Schwarz criterion 10.29027
Log likelihood -121.8942 Hannan-Quinn criter. 10.25421
Durbin-Watson stat 0.021880
EXR at First Difference
Null Hypothesis: D(EXR) is stationary
Exogenous: Constant
Bandwidth: 2 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.367121
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 9.478261
HAC corrected variance (Bartlett kernel) 17.27536
KPSS Test Equation
Dependent Variable: D(EXR)
Method: Least Squares
Date: 01/31/25 Time: 20:18
Sample (adjusted): 2018Q2 2023Q4
Included observations: 23 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 5.000000 0.656376 7.617580 0.0000
R-squared 0.000000 Mean dependent var 5.000000
Adjusted R-squared 0.000000 S.D. dependent var 3.147871
S.E. of regression 3.147871 Akaike info criterion 5.173834
Sum squared resid 218.0000 Schwarz criterion 5.223204
Log likelihood -58.49910 Hannan-Quinn criter. 5.186251
Durbin-Watson stat 1.059633
BTC at Level Form
Null Hypothesis: LBTC is stationary
Exogenous: Constant, Linear Trend
Bandwidth: 3 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.111550
Asymptotic critical values*: 1% level 0.216000
5% level 0.146000
10% level 0.119000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 0.295740
HAC corrected variance (Bartlett kernel) 0.748683
KPSS Test Equation
Dependent Variable: LBTC
Method: Least Squares
Date: 01/31/25 Time: 20:18
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 5.819287 0.737768 7.887692 0.0000
@TREND("2010Q1") 0.088738 0.016749 5.297966 0.0000
R-squared 0.560601 Mean dependent var 9.679388
Adjusted R-squared 0.540629 S.D. dependent var 0.838045
S.E. of regression 0.568001 Akaike info criterion 1.786269
Sum squared resid 7.097756 Schwarz criterion 1.884440
Log likelihood -19.43522 Hannan-Quinn criter. 1.812313
F-statistic 28.06844 Durbin-Watson stat 0.486192
Prob(F-statistic) 0.000026
VBTC at Level Form
Null Hypothesis: LVBTC is stationary
Exogenous: Constant, Linear Trend
Bandwidth: 2 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.142829
Asymptotic critical values*: 1% level 0.216000
5% level 0.146000
10% level 0.119000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 0.032030
HAC corrected variance (Bartlett kernel) 0.063728
KPSS Test Equation
Dependent Variable: LVBTC
Method: Least Squares
Date: 01/31/25 Time: 20:21
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 21.69343 0.242796 89.34855 0.0000
@TREND("2010Q1") 0.020908 0.005512 3.793067 0.0010
R-squared 0.395394 Mean dependent var 22.60293
Adjusted R-squared 0.367912 S.D. dependent var 0.235116
S.E. of regression 0.186926 Akaike info criterion -0.436551
Sum squared resid 0.768711 Schwarz criterion -0.338380
Log likelihood 7.238608 Hannan-Quinn criter. -0.410506
F-statistic 14.38736 Durbin-Watson stat 0.754011
Prob(F-statistic) 0.000998
GDP at Level Form
Null Hypothesis: GDP is stationary
Exogenous: Constant, Linear Trend
Bandwidth: 3 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.087633
Asymptotic critical values*: 1% level 0.216000
5% level 0.146000
10% level 0.119000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 2.836398
HAC corrected variance (Bartlett kernel) 6.456964
KPSS Test Equation
Dependent Variable: GDP
Method: Least Squares
Date: 01/31/25 Time: 20:23
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C -1.996130 2.284802 -0.873656 0.3917
@TREND("2010Q1") 0.092233 0.051872 1.778104 0.0892
R-squared 0.125654 Mean dependent var 2.016003
Adjusted R-squared 0.085911 S.D. dependent var 1.839855
S.E. of regression 1.759049 Akaike info criterion 4.047079
Sum squared resid 68.07356 Schwarz criterion 4.145250
Log likelihood -46.56495 Hannan-Quinn criter. 4.073124
F-statistic 3.161652 Durbin-Watson stat 0.676397
Prob(F-statistic) 0.089213
INF at Level Form
Null Hypothesis: INF is stationary
Exogenous: Constant
Bandwidth: 3 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.651411
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 21.29814
HAC corrected variance (Bartlett kernel) 70.25089
KPSS Test Equation
Dependent Variable: INF
Method: Least Squares
Date: 01/31/25 Time: 20:25
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 16.19952 0.962292 16.83431 0.0000
R-squared 0.000000 Mean dependent var 16.19952
Adjusted R-squared 0.000000 S.D. dependent var 4.714250
S.E. of regression 4.714250 Akaike info criterion 5.979830
Sum squared resid 511.1554 Schwarz criterion 6.028916
Log likelihood -70.75796 Hannan-Quinn criter. 5.992853
Durbin-Watson stat 0.107642
INF at First Difference
Null Hypothesis: D(INF) is stationary
Exogenous: Constant
Bandwidth: 5 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.287979
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 2.093830
HAC corrected variance (Bartlett kernel) 1.255758
KPSS Test Equation
Dependent Variable: D(INF)
Method: Least Squares
Date: 01/31/25 Time: 20:27
Sample (adjusted): 2018Q2 2023Q4
Included observations: 23 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.546280 0.308503 1.770745 0.0905
R-squared 0.000000 Mean dependent var 0.546280
Adjusted R-squared 0.000000 S.D. dependent var 1.479528
S.E. of regression 1.479528 Akaike info criterion 3.663829
Sum squared resid 48.15810 Schwarz criterion 3.713198
Log likelihood -41.13403 Hannan-Quinn criter. 3.676245
Durbin-Watson stat 2.285049
INTR at Level Form
Null Hypothesis: INTR is stationary
Exogenous: Constant, Linear Trend
Bandwidth: 2 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.086865
Asymptotic critical values*: 1% level 0.216000
5% level 0.146000
10% level 0.119000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 1.057093
HAC corrected variance (Bartlett kernel) 2.049152
KPSS Test Equation
Dependent Variable: INTR
Method: Least Squares
Date: 01/31/25 Time: 20:32
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 15.04361 1.394831 10.78526 0.0000
@TREND("2010Q1") -0.271771 0.031667 -8.582246 0.0000
R-squared 0.770007 Mean dependent var 3.221574
Adjusted R-squared 0.759552 S.D. dependent var 2.189983
S.E. of regression 1.073868 Akaike info criterion 3.060067
Sum squared resid 25.37024 Schwarz criterion 3.158238
Log likelihood -34.72080 Hannan-Quinn criter. 3.086111
F-statistic 73.65494 Durbin-Watson stat 0.769150
Prob(F-statistic) 0.000000
TOP at Level Form
Null Hypothesis: TOP is stationary
Exogenous: Constant
Bandwidth: 3 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.688054
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 36.29938
HAC corrected variance (Bartlett kernel) 126.5729
KPSS Test Equation
Dependent Variable: TOP
Method: Least Squares
Date: 01/31/25 Time: 20:36
Sample (adjusted): 2018Q1 2023Q4
Included observations: 24 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 42.22500 1.256278 33.61120 0.0000
R-squared 0.000000 Mean dependent var 42.22500
Adjusted R-squared 0.000000 S.D. dependent var 6.154479
S.E. of regression 6.154479 Akaike info criterion 6.513011
Sum squared resid 871.1850 Schwarz criterion 6.562096
Log likelihood -77.15613 Hannan-Quinn criter. 6.526033
Durbin-Watson stat 0.021408
TOP at First Difference
Null Hypothesis: D(TOP) is stationary
Exogenous: Constant
Bandwidth: 2 (Newey-West automatic) using Bartlett kernel
LM-Stat.
Kwiatkowski-Phillips-Schmidt-Shin test statistic 0.264359
Asymptotic critical values*: 1% level 0.739000
5% level 0.463000
10% level 0.347000
*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)
Residual variance (no correction) 0.231947
HAC corrected variance (Bartlett kernel) 0.454693
KPSS Test Equation
Dependent Variable: D(TOP)
Method: Least Squares
Date: 01/31/25 Time: 20:38
Sample (adjusted): 2018Q2 2023Q4
Included observations: 23 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.760870 0.102679 7.410152 0.0000
R-squared 0.000000 Mean dependent var 0.760870
Adjusted R-squared 0.000000 S.D. dependent var 0.492433
S.E. of regression 0.492433 Akaike info criterion 1.463588
Sum squared resid 5.334783 Schwarz criterion 1.512957
Log likelihood -15.83126 Hannan-Quinn criter. 1.476004
Durbin-Watson stat 0.832274
ARDL
Dependent Variable: EXR
Method: ARDL
Date: 01/31/25 Time: 20:42
Sample (adjusted): 2018Q2 2023Q4
Included observations: 23 after adjustments
Maximum dependent lags: 1 (Automatic selection)
Model selection method: Akaike info criterion (AIC)
Dynamic regressors (1 lag, automatic): LBTC VBTC GDP INF INTR TOP
Fixed regressors: C
Number of models evalulated: 64
Selected Model: ARDL(1, 0, 0, 1, 0, 1, 0)
Variable Coefficient Std. Error t-Statistic Prob.*
EXR(-1) 0.597335 0.094606 6.313902 0.0000
LBTC 3.333460 1.472153 2.264343 0.0413
VBTC -1.44E-09 5.41E-10 -2.665432 0.0194
GDP -0.627838 0.657614 -0.954721 0.3571
GDP(-1) 1.339332 0.697847 1.919236 0.0772
INF 1.351384 0.298378 4.529095 0.0006
INTR -0.097235 1.026029 -0.094769 0.9259
INTR(-1) 2.026941 1.072960 1.889111 0.0814
TOP 2.221578 0.805052 2.759547 0.0162
C 21.13167 24.83557 0.850863 0.4102
R-squared 0.999262 Mean dependent var 410.6522
Adjusted R-squared 0.998751 S.D. dependent var 39.18790
S.E. of regression 1.385152 Akaike info criterion 3.788517
Sum squared resid 24.94240 Schwarz criterion 4.282210
Log likelihood -33.56795 Hannan-Quinn criter. 3.912680
F-statistic 1955.098 Durbin-Watson stat 1.980695
Prob(F-statistic) 0.000000
*Note: p-values and any subsequent tests do not account for model
selection.
COINTEGRATION RESULT
ARDL Long Run Form and Bounds Test
Dependent Variable: D(EXR)
Selected Model: ARDL(1, 0, 0, 1, 0, 1, 0)
Case 2: Restricted Constant and No Trend
Date: 01/31/25 Time: 20:47
Sample: 2010Q1 2023Q4
Included observations: 23
Conditional Error Correction Regression
Variable Coefficient Std. Error t-Statistic Prob.
C 21.13167 24.83557 0.850863 0.4102
EXR(-1)* -0.402665 0.094606 -4.256215 0.0009
LBTC** 3.333460 1.472153 2.264343 0.0413
VBTC** -1.44E-09 5.41E-10 -2.665432 0.0194
GDP(-1) 0.711494 0.917253 0.775679 0.4518
INF** 1.351384 0.298378 4.529095 0.0006
INTR(-1) 1.929706 1.467533 1.314932 0.2113
TOP** 2.221578 0.805052 2.759547 0.0162
D(GDP) -0.627838 0.657614 -0.954721 0.3571
D(INTR) -0.097235 1.026029 -0.094769 0.9259
* p-value incompatible with t-Bounds distribution.
** Variable interpreted as Z = Z(-1) + D(Z).
Levels Equation
Case 2: Restricted Constant and No Trend
Variable Coefficient Std. Error t-Statistic Prob.
LBTC 8.278498 4.701114 1.760965 0.1017
VBTC -3.58E-09 1.51E-09 -2.371753 0.0338
GDP 1.766963 2.027133 0.871656 0.3992
INF 3.356100 0.529176 6.342124 0.0000
INTR 4.792337 3.199724 1.497735 0.1581
TOP 5.517188 1.162976 4.744027 0.0004
C 52.47955 59.01929 0.889193 0.3901
EC = EXR - (8.2785*LBTC -0.0000*VBTC + 1.7670*GDP + 3.3561*INF +
4.7923*INTR + 5.5172*TOP + 52.4796 )
F-Bounds Test Null Hypothesis: No levels relationship
Test Statistic Value Signif. I(0) I(1)
Asymptotic:
n=1000
F-statistic 40.41423 10% 1.99 2.94
k 6 5% 2.27 3.28
2.5% 2.55 3.61
1% 2.88 3.99
Finite Sample:
Actual Sample Size 23 n=35
10% 2.254 3.388
5% 2.685 3.96
1% 3.713 5.326
Finite Sample:
n=30
10% 2.334 3.515
5% 2.794 4.148
1% 3.976 5.691
NORMALITY TEST
HETEROSKEDASTICITY TEST
Heteroskedasticity Test: Breusch-Pagan-Godfrey
F-statistic 0.557454 Prob. F(9,13) 0.8084
Obs*R-squared 6.404639 Prob. Chi-Square(9) 0.6988
Scaled explained SS 2.185412 Prob. Chi-Square(9) 0.9882
Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 01/31/25 Time: 20:51
Sample: 2018Q2 2023Q4
Included observations: 23
Variable Coefficient Std. Error t-Statistic Prob.
C 5.396860 32.10905 0.168079 0.8691
EXR(-1) 0.004272 0.122313 0.034927 0.9727
LBTC 0.200062 1.903296 0.105113 0.9179
VBTC 6.06E-10 7.00E-10 0.865743 0.4023
GDP -0.402113 0.850207 -0.472959 0.6441
GDP(-1) 0.215665 0.902222 0.239037 0.8148
INF 0.010258 0.385763 0.026591 0.9792
INTR -0.227182 1.326517 -0.171262 0.8667
INTR(-1) -0.318998 1.387193 -0.229960 0.8217
TOP -0.237554 1.040824 -0.228237 0.8230
R-squared 0.278463 Mean dependent var 1.084452
Adjusted R-squared -0.221063 S.D. dependent var 1.620622
S.E. of regression 1.790815 Akaike info criterion 4.302240
Sum squared resid 41.69126 Schwarz criterion 4.795933
Log likelihood -39.47576 Hannan-Quinn criter. 4.426402
F-statistic 0.557454 Durbin-Watson stat 2.519057
Prob(F-statistic) 0.808356
AUTOCORRELATION TEST
Breusch-Godfrey Serial Correlation LM Test:
F-statistic 0.097555 Prob. F(1,12) 0.7601
Obs*R-squared 0.185472 Prob. Chi-Square(1) 0.6667
Test Equation:
Dependent Variable: RESID
Method: ARDL
Date: 01/31/25 Time: 20:53
Sample: 2018Q2 2023Q4
Included observations: 23
Presample missing value lagged residuals set to zero.
Variable Coefficient Std. Error t-Statistic Prob.
EXR(-1) 0.000452 0.098082 0.004606 0.9964
LBTC 0.064996 1.540198 0.042200 0.9670
VBTC -1.52E-11 5.63E-10 -0.026979 0.9789
GDP -0.012923 0.682956 -0.018922 0.9852
GDP(-1) -0.080132 0.767554 -0.104399 0.9186
INF 0.017570 0.314381 0.055888 0.9564
INTR -0.098718 1.109577 -0.088969 0.9306
INTR(-1) -0.089298 1.148417 -0.077757 0.9393
TOP -0.074940 0.868345 -0.086302 0.9326
C 2.973270 27.44880 0.108321 0.9155
RESID(-1) -0.121633 0.389429 -0.312338 0.7601
R-squared 0.008064 Mean dependent var -1.52E-14
Adjusted R-squared -0.818549 S.D. dependent var 1.064775
S.E. of regression 1.435887 Akaike info criterion 3.867377
Sum squared resid 24.74126 Schwarz criterion 4.410439
Log likelihood -33.47484 Hannan-Quinn criter. 4.003956
F-statistic 0.009755 Durbin-Watson stat 1.865277
Prob(F-statistic) 1.000000
RAMSEY RESET TEST
Ramsey RESET Test
Equation: UNTITLED
Specification: EXR EXR(-1) LBTC VBTC GDP GDP(-1) INF INTR INTR(-1)
TOP C
Omitted Variables: Squares of fitted values
Value df Probability
t-statistic 1.346252 12 0.2031
F-statistic 1.812394 (1, 12) 0.2031
F-test summary:
Mean
Sum of Sq. df Squares
Test SSR 3.272819 1 3.272819
Restricted SSR 24.94240 13 1.918646
Unrestricted SSR 21.66958 12 1.805798
Unrestricted Test Equation:
Dependent Variable: EXR
Method: ARDL
Date: 01/31/25 Time: 20:54
Sample: 2018Q2 2023Q4
Included observations: 23
Maximum dependent lags: 1 (Automatic selection)
Model selection method: Akaike info criterion (AIC)
Dynamic regressors (1 lag, automatic):
Fixed regressors: C
Variable Coefficient Std. Error t-Statistic Prob.*
EXR(-1) -0.606805 0.899135 -0.674876 0.5125
LBTC -1.446999 3.827393 -0.378064 0.7120
VBTC 5.11E-10 1.54E-09 0.331065 0.7463
GDP 0.734728 1.196413 0.614109 0.5506
GDP(-1) -0.511999 1.532792 -0.334030 0.7441
INF -0.971541 1.749589 -0.555297 0.5889
INTR 0.379030 1.056395 0.358795 0.7260
INTR(-1) -0.863483 2.386045 -0.361889 0.7237
TOP 0.230182 1.672742 0.137607 0.8928
C 336.3964 235.4158 1.428946 0.1785
FITTED^2 0.001989 0.001477 1.346252 0.2031
R-squared 0.999359 Mean dependent var 410.6522
Adjusted R-squared 0.998824 S.D. dependent var 39.18790
S.E. of regression 1.343800 Akaike info criterion 3.734814
Sum squared resid 21.66958 Schwarz criterion 4.277876
Log likelihood -31.95036 Hannan-Quinn criter. 3.871393
F-statistic 1869.730 Durbin-Watson stat 2.082814
Prob(F-statistic) 0.000000
*Note: p-values and any subsequent tests do not account for model
selection.
STABILTITY TEST
GRANGER CAUSALITY
Pairwise Granger Causality Tests
Date: 01/31/25 Time: 21:01
Sample: 2010Q1 2023Q4
Lags: 1
Null Hypothesis: Obs F-Statistic Prob.
BTC does not Granger Cause EXR 23 0.29162 0.5951
EXR does not Granger Cause BTC 0.27934 0.6030
VBTC does not Granger Cause EXR 23 1.57945 0.2233
EXR does not Granger Cause VBTC 0.85047 0.3674
VBTC does not Granger Cause BTC 23 0.69573 0.4141
BTC does not Granger Cause VBTC 0.09899 0.7563
