%This program computes a log-linear approximation to the policy function of a small open economy with an endogenous discount factor (see ``Closing Small Open Economy Models,'' by Stephanie Schmitt-Grohe and Martin Uribe). The reduced form of the model can be written as: %E_t[f(yp,y,xp,x)=0, % %The solution is of the form % %xp = h(x,sigma) + sigma * eta * ep % %y = g(x,sigma) % x is the state (no jump) vector and y is the control (jump) vector. % hx and gx are, respectively, the derivatives of h and g wrt x evaluated at the non-stochastic steady state % %Calls: mendoza91.m gx_hx.m mom.m and ir.m % %(c) Stephanie Schmitt-Grohe and Martin Uribe % %Date November 8, 2001 [fx,fxp,fy,fyp] = mendoza91; [gx,hx] = gx_hx(fy,fx,fyp,fxp); %MOMENTS %Construct Variance/Covariance Matrix of Innovations to State Process varshock = zeros(size(hx,1)); varshock(end,end)=0.0129^2; %Table 6 of Mendoza AER 1991. %Compute Variance/covariance matrix [var_y,var_x] = mom(gx,hx,varshock,0); %First-order autocovariance [var_y1,var_x1] = mom(gx,hx,varshock,1); %Construct a Table with 2nd Moments table(1:6,1) = (diag(var_y(1:6,1:6)).^(1/2))*100; table(1:6,2) = diag(var_y1(1:6,1:6)) ./ diag(var_y(1:6,1:6)); table(1:6,3) = var_y(1:6,1) ./ (diag(var_y(1:6,1:6)).^(1/2)) /var_y(1,1)^(1/2) %Impulse responses x0 = zeros(size(hx,1),1); x0(end) = 1; IR=ir(gx,hx,x0,10); t=(0:size(IR,1)-1)'; y = IR(:,1); c = IR(:,2); i = IR(:,3); h = IR(:,4); tby = IR(:,5); cay = IR(:,6); subplot(3,2,1) plot(t,y) title('Output') subplot(3,2,2) plot(t,c) title('Consumption') subplot(3,2,3) plot(t,i) title('Investment') subplot(3,2,4) plot(t,h) title('Hours') subplot(3,2,5) plot(t,tby) title('Trade Balance / GDP') subplot(3,2,6) plot(t,cay) title('Current Account / GDP') shg