future price

(1)

Appendix contents

1. Model specifications and forecasts ... 37

2. Correlations ... 41

3. Rebel against 2^nd future price... 41

4. Classification process ... 42

(2)

onvergence achieved after 26 iterations riance backcast: ON ICE = C(1) + C(2)*EXPORT(-1) + [AR(1)=C(3)] ARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C1487,0646,9331,6890,000 EXPORT(-1)-0,1930,1144-1,6830,092 AR(1)0,9440,020845,4540,000 Variance Equation C3190,82145,5921,9170,000 RESID(-1)^20,0420,01004,1880,000 GARCH(-1)-1,0230,0018-572,1060,000 R-squared0,902 Mean dependent var1504,660 Adjusted R-squared0,900 S.D. dependent var109,603 S.E. of regression34,604 Akaike info criterion9,941 Sum squared resid355637 Schwarz criterion10,014 Log likelihood-1500,029 F-statistic546,547 Durbin-Watson stat1,867 Prob(F-statistic)0,000 Inverted AR Roots0,94010 day forecastt-Statistic Prob.20 day forecastt-Statistic Prob. Augmented Dickey-Fuller test statistic-2,35400,1754Augmented Dickey-Fuller test statistic-2,65930,0984 Test critical values:1% level-3,8085Test critical values:1% level-3,8085 5% level-3,02075% level-3,0207 10% level-2,650410% level-2,6504

Export 1350

1400

1450

1500

1550

1600

1650

1700 588589590591592593594595596597 FORECAST_EXPORT

Forecast: FORECAST_EXPORT Actual: PRICE Forecast sample: 588 597 Included observations: 10 Root Mean Squared Error 23.99302 Mean Absolute Error 15.48778 Mean Abs. Percent Error 1.037170 Theil Inequality Coefficient 0.008040 Bias Proportion 0.012128 Variance Proportion 0.006530 Covariance Proportion 0.981342 500

1000

1500

2000

2500

3000

3500 588589590591592593594595596597 Forecast of Variance

1300

1400

1500

1600

1700

1800 578580582584586588590592594596 FORE20_EXPORT

Forecast: FORE20_EXPORT Actual: PRICE Forecast sample: 578 597 Included observations: 20 Root Mean Squared Error 32.20201 Mean Absolute Error 21.45369 Mean Abs. Percent Error 1.415981 Theil Inequality Coefficient 0.010635 Bias Proportion 0.047162 Variance Proportion 0.131635 Covariance Proportion 0.821203 900

1000

1100

1200

1300

1400

1500

1600

onvergence achieved after 35 iterations ariance backcast: ON ARCH = C(6) + C(7)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C1495,0832,6345,8150,000 REBEL(-2)0,4700,16482,8520,004 SUN(-5)-1,2210,5931-2,0590,039 AR(1)0,9850,026237,5640,000 AR(6)-0,0490,0228-2,1510,032 Variance Equation C2271,570,161388,1790,000 GARCH(-1)-0,9960,0041-245,8190,000 R-squared0,906 Mean dependent var1504,660 Adjusted R-squared0,904 S.D. dependent var109,603 S.E. of regression33,881 Akaike info criterion9,886 Sum squared resid339779.1 Schwarz criterion9,972 Log likelihood-1490,714 F-statistic477,409 Durbin-Watson stat1,958 Prob(F-statistic)0,000 Inverted AR Roots .90 .71 .11-.53i .11+.53i10 day forecastt-Statistic Prob.20 day forecastt-Statistic Prob. -.42+.28i -.42-.28iAugmented Dickey-Fuller test statistic-2,47460,1483Augmented Dickey-Fuller test statistic-2,56920,1155 Test critical values:1% level-3,8085Test critical values:1% level-3,8085 5% level-3,02075% level-3,0207 10% level-2,650410% level-2,6504

Rebel, Sun 1360

1400

1440

1480

1520

1560

1600

1640 588589590591592593594595596597 FORECAST_REBEL_SUN

Forecast: FORECAST_REBEL_SUN Actual: PRICE Forecast sample: 588 597 Included observations: 10 Root Mean Squared Error 24.54000 Mean Absolute Error 16.99973 Mean Abs. Percent Error 1.140984 Theil Inequality Coefficient 0.008223 Bias Proportion 0.013036 Variance Proportion 0.001362 Covariance Proportion 0.985602 1000

1050

1100

1150

1200

1250

1300

1400

1500

1600

1700

1800 578580582584586588590592594596 FORE20_REBEL_SUN

Forecast: FORE20_REBEL_SUN Actual: PRICE Forecast sample: 578 597 Included observations: 20 Root Mean Squared Error 34.08770 Mean Absolute Error 22.73624 Mean Abs. Percent Error 1.505450 Theil Inequality Coefficient 0.011249 Bias Proportion 0.073400 Variance Proportion 0.111015 Covariance Proportion 0.815585 950

1000

1050

1100

1150

1200

1250

1. Model specifications and forecasts

(3)

Convergence achieved after 26 iterations Variance backcast: ON PRICE = C(1) + C(2)*IMPORT(-3) + [AR(1)=C(3)] GARCH = C(4) + C(5)*GARCH(-1) + C(6)*GARCH(-2) CoefficientStd. Errorz-StatisticProb. C1548,2650,7330,5220,000 IMPORT(-3)-1,1260,6705-1,6800,093 AR(1)0,9640,021644,7160,000 Variance Equation C4225,65270,2015,6390,000 GARCH(-1)-0,8860,0045-195,9220,000 GARCH(-2)-1,0190,0091-112,1390,000 R-squared0,902 Mean dependent var1504,660 Adjusted R-squared0,901 S.D. dependent var109,603 S.E. of regression34,566 Akaike info criterion9,893 Sum squared resid354858 Schwarz criterion9,967 Log likelihood-1492,791 F-statistic547,876 Durbin-Watson stat1,882 Prob(F-statistic)0,000 Inverted AR Roots0,96010 day forecastt-Statistic Prob.20 day forecastt-Statistic Prob. Augmented Dickey-Fuller test statistic-2,39850,1650Augmented Dickey-Fuller test statistic-2,56760,1158 Test critical values:1% level-3,8085Test critical values:1% level-3,8085 5% level-3,02075% level-3,0207 10% level-2,650410% level-2,6504

Import 1320

1360

1400

1440

1480

1520

1560

1600

1640

1680 588589590591592593594595596597 FORECAST_IMPORT

Forecast: FORECAST_IMPORT Actual: PRICE Forecast sample: 588 597 Included observations: 10 Root Mean Squared Error 23.53870 Mean Absolute Error 15.67390 Mean Abs. Percent Error 1.051223 Theil Inequality Coefficient 0.007883 Bias Proportion 0.034361 Variance Proportion 0.007065 Covariance Proportion 0.958574 0

1000

2000

3000

4000

1300

1400

1500

1600

1700

1800 578580582584586588590592594596 FORE20_IMPORT

Forecast: FORE20_IMPORT Actual: PRICE Forecast sample: 578 597 Included observations: 20 Root Mean Squared Error 33.21730 Mean Absolute Error 21.86131 Mean Abs. Percent Error 1.443388 Theil Inequality Coefficient 0.010958 Bias Proportion 0.097735 Variance Proportion 0.118089 Covariance Proportion 0.784176 0

400

800

1200

1600

2000

2400 578580582584586588590592594596 Forecast of Variance Convergence achieved after 30 iterations Variance backcast: ON PRICE = C(1) + C(2)*FIGHT(-3) + [AR(1)=C(3)] GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1) CoefficientStd. Errorz-StatisticProb. C1461,7714,9297,9620,000 FIGHT(-3)-1,4500,8803-1,6480,099 AR(1)0,9230,0061150,6450,000 Variance Equation C2318,2813,82167,8010,000 RESID(-1)^20,0150,000436,5400,000 GARCH(-1)-1,0200,0005-1896,4670,000 R-squared0,902 Mean dependent var1504,660 Adjusted R-squared0,900 S.D. dependent var109,603 S.E. of regression34,666 Akaike info criterion9,907 Sum squared resid356913 Schwarz criterion9,981 Log likelihood-1494,976 F-statistic544,380 Durbin-Watson stat1,796 Prob(F-statistic)0,000 Inverted AR Roots0,92010 day forecastt-Statistic Prob.20 day forecastt-Statistic Prob. Augmented Dickey-Fuller test statistic-2,21890,2114Augmented Dickey-Fuller test statistic-2,66230,0979 Test critical values:1% level-3,8085Test critical values:1% level-3,8085 5% level-3,02075% level-3,0207 10% level-2,650410% level-2,6504

Fight 1360

1400

1440

1480

1520

1560

1600

1640 588589590591592593594595596597 FORECAST_FIGHT

Forecast: FORECAST_FIGHT Actual: PRICE Forecast sample: 588 597 Included observations: 10 Root Mean Squared Error 23.66338 Mean Absolute Error 15.65388 Mean Abs. Percent Error 1.049043 Theil Inequality Coefficient 0.007934 Bias Proportion 0.001604 Variance Proportion 0.012939 Covariance Proportion 0.985457 800

1200

1600

2000

2400

1300

1400

1500

1600

1700

1800 578580582584586588590592594596 FORE20_FIGHT

Forecast: FORE20_FIGHT Actual: PRICE Forecast sample: 578 597 Included observations: 20 Root Mean Squared Error 32.61396 Mean Absolute Error 21.04653 Mean Abs. Percent Error 1.389179 Theil Inequality Coefficient 0.010756 Bias Proportion 0.118512 Variance Proportion 0.103150 Covariance Proportion 0.778338 0

1000

2000

3000

4000