// Test for var(log) in a deterministic context var(log) c k; varexo x; parameters alph gam delt bet aa; alph=0.5; gam=0.5; delt=0.02; bet=0.05; aa=0.5; model; c + k - aa*x*k(-1)^alph - (1-delt)*k(-1); c^(-gam) - (1+bet)^(-1)*(aa*alph*x(+1)*k^(alph-1) + 1 - delt)*c(+1)^(-gam); end; initval; x = 1; k = ((delt+bet)/(1.0*aa*alph))^(1/(alph-1)); c = aa*k^alph-delt*k; end; steady; check; shocks; var x; periods 1; values 1.2; end; perfect_foresight_setup(periods=200); perfect_foresight_solver; idx_c = strmatch('c', M_.endo_names); idx_log_c = strmatch('LOG_c', M_.endo_names); if isempty(idx_log_c) error('Log-transformed variable not created') end if max(abs(oo_.endo_simul(idx_log_c, :) - log(oo_.endo_simul(idx_c, :)))) > 1e-7 error('Transformation not correctly performed') end S = load('../deterministic_simulations/ramst/Output/ramst_results.mat'); if max(abs(oo_.endo_simul(idx_c, :) - S.oo_.endo_simul(idx_c, :))) > 1e-7 error('Result differs from non-transformed model') end