Known bugs present in the current stable version

This page documents known bugs in Dynare stable.

• If the guess value given to the steady command is such that the residuals are all below tolerance, except some that are NaN, then this guess value is incorrectly accepted as the solution to the steady state problem (bug fixed in 3a789ca7)
• Homotopy with extended_path at order=0 is broken (bug fixed in 6ed90b3d)
• The command line option parallel_use_psexec will be ignored (bug fixed in 4fa61014)
• With the bytecode option of model block, using the operators abs(), cbrt() and sign() will lead to a crash (bug fixed in 1622c74a)
• With the use_dll option of model block, the expression sign(x) evaluates to ±1 instead of 0 if x=0 (bug fixed in preprocessor@b1e48842)
• In a stochastic context, results can be incorrect if an endogenous with a lead ⩾ 2 or an exogenous with a lead ⩾ 1 appears in the argument(s) of a call to a (nonlinear) external function (bug fixed in preprocessor@712b11a0)
• Ramsey steady state computation can fail if expectation or diff operator is present in the model (bug fixed in 674ebce8)
• A crash can occur if some external function call is present in an auxiliary variable (bug fixed in preprocessor@34c37cfd)
• The endogenous_prior-option may erroneously display a warning message about missing observations (bug fixed in a85a00bf)
• model_comparison will crash if the .mod file name has less than four characters (bug fixed in 183e6fbb)
• shock_decomposition will overwrite previously stored smoother results (bug fixed in 9e53bd99)
• x13 does not handle missing values, particularly at the beginning of a series (bug fixed in dseries@95a011a7 and dseries@c0786192)
• x13 would occasionally crash under Windows with segmentation violations (bug fixed in 39c2cba1)
• OccBin: estimation will crash if a previous shocks(surprise)-simulation was conducted (bug fixed in 37eb1097)
• Theoretical moments at order=2 with pruning do not contain unconditional and conditional variance decomposition (regression from 4.6; fixed in 41ac891f and c063d536; not backported to 5.x)
• Forecasts are only based on the DSGE-model, not the DSGE-VAR (bug not yet fixed; see #819)
• The bvar_irf command would ignore the SquareRoot option and instead employ a Cholesky decomposition (bug not yet fixed, see !2175 (closed))
• The contemp_reduced_form-option of the sbvar-command is broken (bug not yet fixed, see #1716)