diff --git a/matlab/extended_path.m b/matlab/extended_path.m
index 676dcefd1..926c79d02 100644
--- a/matlab/extended_path.m
+++ b/matlab/extended_path.m
@@ -86,7 +86,7 @@ verbose = 0;
 t = 0;
 new_draw = 1;
 
-perfect_foresight_simulation([],[],[]);
+perfect_foresight_simulation();
 
 while (t<=sample_size)
     t = t+1;
diff --git a/matlab/perfect_foresight_simulation.m b/matlab/perfect_foresight_simulation.m
index fd50f8759..45af9b0f5 100644
--- a/matlab/perfect_foresight_simulation.m
+++ b/matlab/perfect_foresight_simulation.m
@@ -1,4 +1,4 @@
-function info = perfect_foresight_simulation(compute_linear_solution,steady_state, init_flag)
+function info = perfect_foresight_simulation(compute_linear_solution,steady_state)
 % Performs deterministic simulations with lead or lag on one period
 %
 % INPUTS
@@ -37,9 +37,9 @@ function info = perfect_foresight_simulation(compute_linear_solution,steady_stat
 
 global M_ options_ it_ oo_
 
-persistent lead_lag_incidence dynamic_model ny nyp nyf nrs nrc iyf iyp isp is isf isf1 iz icf ghx
+persistent lead_lag_incidence dynamic_model ny nyp nyf nrs nrc iyf iyp isp is isf isf1 iz icf ghx iflag
 
-if nargin==3% Initialization of the persistent variables.
+if ~nargin && isempty(iflag)% Initialization of the persistent variables.
     lead_lag_incidence = M_.lead_lag_incidence; 
     dynamic_model = [M_.fname '_dynamic'];
     ny   = size(oo_.endo_simul,1); 
@@ -54,8 +54,9 @@ if nargin==3% Initialization of the persistent variables.
     isf  = iyf+nyp; 
     isf1 = (nyp+ny+1):(nyf+nyp+ny+1);     
     iz   = 1:(ny+nyp+nyf);
-    icf  = 1:size(iyf,2); 
+    icf  = 1:size(iyf,2);
     info = [];
+    iflag = 1;
     return
 end