This is machine translation

Translated by Microsoft
Mouseover text to see original. Click the button below to return to the English version of the page.

Note: This page has been translated by MathWorks. Click here to see
To view all translated materials including this page, select Country from the country navigator on the bottom of this page.

Forecast VAR Model Using Monte Carlo Simulation

This example shows how to use Monte Carlo simulation via simulate to forecast a VAR model.

simulate enables you to generate simulations of time series based on your model. If you have a trustworthy VAR model object, you can use these simulations as sample forecasts.

simulate requires:

  • A model (EstMdl in what follows)

  • The number of periods for the forecast (numobs in what follows)

simulate optionally takes:

  • An exogenous data series

  • A presample time series (Y(end-3:end,:) in what follows)

  • Future sample responses for conditional simulation

  • The number of realizations, or paths, to simulate (2000 in what follows)

Load the Data_USEconModel data set. This example uses two time series: the logarithm of real GDP, and the real 3-month T-bill rate, both differenced to be approximately stationary. For illustration, a VAR(4) model describes the time series.

load Data_USEconModel
DEF = log(DataTable.CPIAUCSL);
GDP = log(DataTable.GDP);
rGDP = diff(GDP - DEF); % Real GDP is GDP - deflation
TB3 = 0.01*DataTable.TB3MS;
dDEF = 4*diff(DEF); % Scaling
rTB3 = TB3(2:end) - dDEF; % Real interest is deflated
Y = [rGDP,rTB3];

Fit a VAR(4) model specification.

Mdl = varm(2,4);
Mdl.SeriesNames = {'Transformed real GDP','Transformed real 3-mo T-bill rate'};
EstMdl = estimate(Mdl,Y);

Define the forecast horizon.

numobs = 21;
FDates = dateshift(DataTable.Time(end),'end','quarter',1:numobs);

Simulate the model for numobs steps ahead, and generate 2000 paths. Specify presample observations from the end of the data.

rng(1); %For reproducibility
Ysim = simulate(EstMdl,numobs,'Y0',Y(end-3:end,:),'NumPaths',2000);

Calculate the mean and standard deviation of the simulated series:

Ymean = mean(Ysim,3); % Calculate means
Ystd = std(Ysim,0,3); % Calculate std deviations

Plot the means +/- 1 standard deviation for the simulated series:

plot([DataTable.Time(end) FDates],[Y(end,1);Ymean(:,1)],'r')
plot([DataTable.Time(end) FDates],[Y(end,1);Ymean(:,1)]+[0;Ystd(:,1)],'b')
plot([DataTable.Time(end) FDates],[Y(end,1);Ymean(:,1)]-[0;Ystd(:,1)],'b')
title('Transformed real GDP')
plot([DataTable.Time(end) FDates],[Y(end,2);Ymean(:,2)],'r')
plot([DataTable.Time(end) FDates],[Y(end,2);Ymean(:,2)]+[0;Ystd(:,2)],'b')
plot([DataTable.Time(end) FDates],[Y(end,2);Ymean(:,2)]-[0;Ystd(:,2)],'b')
title('Transformed real 3-mo T-bill rate')

See Also



Related Topics