In a stochastic context, Dynare computes one or several simulations corresponding to a random draw of the shocks.

The main algorithm for solving stochastic models relies on a Taylor approximation, up to second order, of the solution function (see Judd (1996), Collard and Juillard (2001a, 2001b), and Schmitt-Grohé and Uríbe (2004)). The details of the Dynare implementation of the first order solution are given in Villemot (2011). Such a solution is computed using the stoch_simul command.

Dynare commands

stoch_simul

Command: `stoch_simul;

Command: `stoch_simul (OPTIONS...);

Solves a stochastic (i.e. rational expectations) model, using perturbation techniques.

More precisely, stoch_simul computes a Taylor approximation of the model around the deterministic steady state and solves of the the decision and transition functions for the approximated model. Using this, it computes impulse response functions and various descriptive statistics (moments, variance decomposition, correlation and autocorrelation coefficients). For correlated shocks, the variance decomposition is computed as in the VAR literature through a Cholesky decomposition of the covariance matrix of the exogenous variables. When the shocks are correlated, the variance decomposition depends upon the order of the variables in the varexo command.

The IRFs are computed as the difference between the trajectory of a variable following a shock at the beginning of period 1 and its steady state value. More details on the computation of IRFs can be found at https://archives.dynare.org/DynareWiki/IrFs.

Variance decomposition, correlation, autocorrelation are only displayed for variables with strictly positive variance. Impulse response functions are only plotted for variables with response larger than $10^{-10}$.

Variance decomposition is computed relative to the sum of the contribution of each shock. Normally, this is of course equal to aggregate variance, but if a model generates very large variances, it may happen that, due to numerical error, the two differ by a significant amount. Dynare issues a warning if the maximum relative difference between the sum of the contribution of each shock and aggregate variance is larger than 0.01%.

The covariance matrix of the shocks is specified with the shocks command (see shocks-exo).

Options

• ar = INTEGER

Order of autocorrelation coefficients to compute. Default: 5

• irf = INTEGER

Number of periods on which to compute the IRFs. Setting irf=0 suppresses the plotting of IRFs. Default: 40.

• nonstationary: declares the model as nonstationary.

• noprint: don't print the results

• order = INTEGER

Order of Taylor approximation. Note that for third order and above, the k_order_solver option is implied and only empirical moments are available (you must provide a value for periods option). Default: 2

• periods = INTEGER

If different from zero, empirical moments will be computed instead of theoretical moments. The value of the option specifies the number of periods to use in the simulations. Values of the initval block, possibly recomputed by steady, will be used as starting point for the simulation. The simulated endogenous variables are made available to the user in Julia variable context.results.model_results[1].simulation. Default: 0.

• dr = OPTION

Determines the method used to compute the decision rule. Possible values for OPTION are:

default

Uses the default method to compute the decision rule based on the generalized Schur decomposition (see Villemot (2011) for more information).

cycle_reduction

Uses the cycle reduction algorithm to solve the polynomial equation for retrieving the coefficients associated to the endogenous variables in the decision rule. This method is faster than the default one for large scale models.

Default value is default.

Output

The derivatives of the approximated solution function are availabe in the vector of matrices context.results.model_results[1].solution_derivatives. The first element contains the matrix of first order derivatives. The second element, the matrix of second order derivatives.

The matrix of first order derivatives is a $n x (n_s + n_x + 1)$ matrix where n is the number of endogenous variables, $n_s$, the number of state variables (variables appearing in the model with a lag), and $n_x$, the number of exogenous variables. An element of this matrix is

\[\begin{align*} X_{i,j} &= \frac{\partial g_i}{\partial y_j},\;\;j=1,\ldots,n_s\\ X_{i,n_s+j} &= \frac{\partial g_i}{\partial x_j},\;\;j=1,\ldots,n_x\\ X_{i,n_s+n_k+1} &= \frac{\partial g_i}{\partial \sigma} = 0 \end{align*}\]

where $g_i$ is the solution function for variable i, y, the vector of endogenous variables, x, the vector en exogenous variables and $\sigma$ the stochastic scale of the model. Note that at first order, this derivative is alwasy equal to zero.

The matrix of second order derivatives is $n \times n^2$ matrix where each column contains derivatives with respect to a pair of endogenous variables

• eigenvalues::Vector{Complex{Float64}}
• g1::Matrix{Float64} # full approximation
• gs1::Matrix{Float64} # state transition matrices: states x states
• hs1::Matrix{Float64} # states x shocks
• gns1::Matrix{Float64} # non states x states
• hns1::Matrix{Float64} # non states x shocsks
• g1_1::SubArray{Float64, 2, Matrix{Float64}, Tuple{Base.Slice{Base.OneTo{Int}}, UnitRange{Int}}, true} # solution first order derivatives w.r. to state variables
• g1_2::SubArray{Float64, 2, Matrix{Float64}, Tuple{Base.Slice{Base.OneTo{Int}}, UnitRange{Int}}, true} # solution first order derivatives w.r. to current exogenous variables
• endogenous_variance::Matrix{Float64}
• stationary_variables::Vector{Bool}

Example 1

shocks;
var e;
stderr 0.0348;
end;

stoch_simul;

Performs the simulation of the 1st-order approximation of a model with a single stochastic shock e, with a standard error of 0.0348.

Example 2

    stoch_simul(irf=60);

Performs the simulation of a model and displays impulse response functions on 60 periods.

Julia function

localapproximation!(; context::Context=context, display = true,
                    dr_algo::String = "GS", irf::Int = 40,
                    LRE_options = LinearRationalExpectationsOptions(),
                    nar::Int = 5, nonstationary::Bool = false,
                    order::Int = 1, periods::Int = 0 )

First-order approximation

The approximation has the stylized form:

\[y_t = y^s + A \phi(y_{t-1}) + B u_t\]

where $y^s$ is the steady state value of $y$ and $\phi(y_{t-1})=y_{t-1}-y^s$. Matrices of coefficients $A$ and $B$ are computed by Dynare.

Second-order approximation

The approximation has the form:

\[y_t = y^s + 0.5 \Delta^2 + A \phi(y_{t-1}) + B u_t + 0.5 C (\phi(y_{t-1})\otimes \phi(y_{t-1})) + 0.5 D (u_t \otimes u_t) + E (\phi(y_{t-1}) \otimes u_t)\]

where $y^s$ is the steady state value of $y$, $\phi(y_{t-1})=y_{t-1}-y^s$, and $\Delta^2$ is the shift effect of the variance of future shocks. Matrices of coefficients $A$, $B$, $C$, $D$ and $E$ are computed by Dynare.