Select Page

Code For Only FTSE and Rolling Window Case Solution

%%%%%%%%%%%%% making all variables equal to IVI360:

%%FTSE 100

f = xlsread('data64.xlsx','C3:C94')./ 100;

%values of All Parameters are no equal to the daily data so we are

[~,datenums] = xlsread('data64.xlsx',2,'A7:A1918');

[~,datemonth] = datevec(datenums);

element = size(datenums,1);

f100 = NaN(element,1);

gin= 1;

for t = 1:element

if t > 1 && datemonth(t) ~= datemonth(t-1)

gin=gin+1;

if gin> length(f)

break

end

end

f100(t)=f(gin);

end

fprintf('Table For Ftse Only\n')

[estParams,EstParamCov,Variance,LongRunVar] = GarchMidas(subsample1,'Period',period,'NumLags',16,'X',f100,'ThetaM',1);

subplot(2,1,1);

plot(year(seq),sqrt(252*Variance(seq)),'g','LineWidth',1);

hold on

plot(year(seq),sqrt(252*LongRunVar(seq)),'b--','LineWidth',2);

legend('TV','SV','Location','SouthEast')

xlim([2000,2007])

ylim([0,b])

title('Ftse W.P')

hold off

fprintf('Table For Ftse with rolling window\n')

[estParams,EstParamCov,Variance,LongRunVar] = GarchMidas(subsample1,'Period',period,'NumLags',16,'X',f100,'ThetaM',1,'RollWindow',1);

subplot(2,1,2);

plot(year(seq),sqrt(252*Variance(seq)),'g','LineWidth',1);

hold on

plot(year(seq),sqrt(252*LongRunVar(seq)),'b--','LineWidth',2);

legend('TV','SV','Location','SouthEast')

xlim([2000,2007])

ylim([0,b])

title('Ftse W.P with rolling window')

hold off

fprintf ('\nTable for Garch Midas With Ftse in sample\n');

% In-sample forecast validation

GarchMidas(subsample1,'Period',period,'NumLags',numLags,'estSample',1900);

fprintf ('\nTable for Garch Midas with Ftse out of sample\n');

% Out-of-sample forecast

estParams = GarchMidas(subsample1,'Period',period,'NumLags',numLags);

nForecast = 10;

yBig = [subsample1;0];

for t = 1:nForecast

[~,~,Variance,LongRunVar] = GarchMidas(yBig,'Period',period,'NumLags',numLags,'Params',estParams);

yPseudo = estParams(1) + sqrt(Variance(end));

yBig = [yBig(1:end-1);yPseudo;0];

end

VarianceForecast = Variance(element+1:element+nForecast);

LongRunVarForecast = LongRunVar(element+1:element+nForecast);

rolling window

rolling window

 

Tables:

Table For Ftse Only

Method: Maximum likelihood

Sample size: 1912

Adjusted sample size: 1496

Logarithmic  likelihood:      2866.62

Akaike   info criterion:     -5721.24

Bayesian info criterion:     -5687.91

Coeff        StdErr        tStat       Prob

mu         0.13528    0.00042046       321.75          0

alpha      0.84698       0.13207       6.4129          0

beta        0.1476      0.091204       1.6184    0.10559

theta      -4.3961        71.359    -0.061605    0.95088

w            48.18         515.4     0.093481    0.92552

m        0.0068454       0.10985     0.062315    0.95031

RMSE of one-step variance forecast (period 1 to 1912): 6.086e-03.

 

 

 

Table For Ftse with rolling window

Method: Maximum likelihood

Sample size: 1912

Adjusted sample size: 1496

Logarithmic  likelihood:      2867.51

Akaike   info criterion:     -5723.01

Bayesian info criterion:     -5689.68

Coeff       StdErr        tStat        Prob

mu         0.13448    0.0004475       300.52           0

alpha      0.85106      0.14013       6.0731           0

beta       0.14154      0.09676       1.4628     0.14352

theta      -20.641        248.8    -0.082964     0.93388

w           1.0013      0.52979         1.89    0.058763

m        0.0063926     0.076475     0.083591     0.93338

RMSE of one-step variance forecast (period 1 to 1912): 6.155e-03.

 

 

Table for Garch Midas With Ftse in sample

Method: Maximum likelihood

Sample size: 1900

Adjusted sample size: 1484

Logarithmic  likelihood:      2856.26

Akaike   info criterion:     -5700.51

Bayesian info criterion:     -5667.21

Coeff        StdErr       tStat        Prob

mu        0.13444    0.00045203     297.41             0

alpha     0.84968       0.14054     6.0456             0

beta      0.14244      0.097251     1.4647       0.14301

theta    0.042136       0.24164    0.17437       0.86157

w           1.001       0.22686     4.4125    1.0222e-05

m        0.053367       0.30785    0.17335       0.86238

RMSE of one-step variance forecast (period 1 to 1900): 6.278e-03.

RMSE of one-step variance forecast (period 1901 to 1912): 3.824e-03.

 

 

Table for Garch Midas with Ftse out of sample

Method: Maximum likelihood

Sample size: 1912

Adjusted sample size: 1496

Logarithmic  likelihood:      2867.83

Akaike   info criterion:     -5723.67

Bayesian info criterion:     -5690.33

Coeff       StdErr       tStat        Prob

mu        0.13474    0.0004485     300.43             0

alpha     0.85086      0.13935      6.106             0

beta      0.14346     0.096542      1.486       0.13729

theta    0.048784       0.3847    0.12681       0.89909

w           1.001      0.23446     4.2694    1.9603e-05

m        0.063143      0.49982    0.12633       0.89947

RMSE of one-step variance forecast (period 1 to 1912): 6.120e-03.

RMSE of one-step variance forecast (period 1 to 1913): 6.126e-03.

RMSE of one-step variance forecast (period 1 to 1914): 6.125e-03.

RMSE of one-step variance forecast (period 1 to 1915): 6.123e-03.

RMSE of one-step variance forecast (period 1 to 1916): 6.121e-03.

RMSE of one-step variance forecast (period 1 to 1917): 6.120e-03.

RMSE of one-step variance forecast (period 1 to 1918): 6.118e-03.

RMSE of one-step variance forecast (period 1 to 1919): 6.116e-03.

RMSE of one-step variance forecast (period 1 to 1920): 6.115e-03.

RMSE of one-step variance forecast (period 1 to 1921): 6.113e-03.

RMSE of one-step variance forecast (period 1 to 1922): 6.112e-03.............................

This is just a sample partial case solution. Please place the order on the website to order your own originally done case solution.

Share This