simulation_functions: Low-level MDCEV simulation functions

Description

Compiled C++ routines that underlie mdcev.sim and GenerateMDCEVData. These functions are exported for advanced users and test code; most users should call the higher-level wrappers instead.

DrawError_rng: Draw preference error terms for one individual using the MLHS or independent draw approach.
MarshallianDemand: Compute Marshallian (uncompensated) demand for one individual given marginal utilities at zero.
HicksianDemand: Compute Hicksian (compensated) demand for one individual given a reference utility level.
ComputeUtilJ: Evaluate the KT utility function for one individual.
CalcWTP_rng: Simulate welfare (WTP) across nerrs error draws for one individual under a price policy.
CalcMarshallianDemand_rng: Simulate Marshallian demand across nerrs error draws for one individual under a price policy.

Usage

DrawError_rng(quant_num, quant_j, price_j, psi_j, phi_j, gamma_j,
              alpha, scale, model_num, nalts, nerrs, cond_error,
              draw_mlhs, base_rng__, pstream__)
MarshallianDemand(income, price, MUzero, phi, gamma, alpha, nalts,
                  algo_gen, model_num, tol_e, max_loop, pstream__)
HicksianDemand(util, price, MUzero, phi, gamma, alpha, nalts,
               algo_gen, model_num, tol_l, max_loop, pstream__)
ComputeUtilJ(income, quant_j, price_j, psi, phi_j, gamma_j, alpha,
             nalts, model_num, pstream__)
CalcWTP_rng(income, quant_j, price, price_p_policy, psi_p_sims,
            phi_p_sims, psi_sims, phi_sims, gamma_sims, alpha_sims,
            scale_sims, nerrs, cond_error, draw_mlhs, algo_gen,
            model_num, price_change_only, tol, max_loop, base_rng__,
            pstream__)
CalcMarshallianDemand_rng(income, quant_j, price, price_p_policy,
                          psi_p_sims, phi_p_sims, psi_sims, phi_sims,
                          gamma_sims, alpha_sims, scale_sims, nerrs,
                          cond_error, draw_mlhs, algo_gen, model_num,
                          price_change_only, tol, max_loop, base_rng__,
                          pstream__)

Value

DrawError_rng returns a matrix of error draws with nerrs

rows and nalts + 1 columns.

MarshallianDemand and HicksianDemand return a numeric vector of length nalts + 1 holding the numeraire quantity followed by the non-numeraire quantities.

ComputeUtilJ returns a scalar utility value.

CalcWTP_rng and CalcMarshallianDemand_rng return a list with one element per policy scenario, each holding the simulated welfare measure or demand averaged over the nerrs error draws.

Arguments

base_rng__: External pointer to the Boost RNG created by rmdcev_get_rng.
pstream__: External pointer to the output stream created by rmdcev_get_stream.
model_num: Integer model type: 1 = gamma, 2 = alpha, 3 = hybrid, 4 = hybrid0, 5 = kt_ee, 6 = gamma1.
income: Individual income (scalar).
quant_j: Vector of non-numeraire quantities.
quant_num: Numeraire quantity consumed.
price_j: Vector of non-numeraire prices.
price: Full price vector (numeraire first).
price_p_policy: Matrix of counterfactual prices, one row per policy scenario.
psi_j, psi: Psi utility parameters.
phi_j, phi: Phi parameters (kt_ee model).
gamma_j, gamma: Gamma satiation parameters.
alpha: Alpha parameters.
scale: Scale parameter.
psi_sims, phi_sims, gamma_sims, alpha_sims, scale_sims: Lists of parameter draws, one element per simulation, for the baseline scenario.
psi_p_sims, phi_p_sims: Lists of psi and phi parameter draws for the policy scenarios.
nalts: Number of non-numeraire alternatives.
nerrs: Number of error draws.
cond_error: 1 = conditional error draws; 0 = unconditional.
draw_mlhs: 1 = Modified Latin Hypercube Sampling; 0 = iid.
algo_gen: Algorithm: 0 = hybrid (fast, requires alpha = 0); 1 = general bisection.
price_change_only: 1 = only prices change across policies, so the baseline psi and phi draws are reused; 0 = psi and phi also change across policies.
MUzero: Vector of marginal utilities at zero consumption.
util: Reference utility level (Hicksian demand).
tol, tol_e, tol_l: Convergence tolerance for bisection.
max_loop: Maximum bisection iterations.

Description

Usage

Value

Arguments

See Also