| Type: | Package |
| Title: | Identification, Estimation and Inference Based on Structural Error Projection |
| Version: | 0.1.0 |
| Description: | Estimation and inference for regression models with endogenous regressors using a semiparametric projection approach. Instrumental variables are constructed internally from observed regressors by projecting out a space of basis functions used to represent the conditional mean of the structural error. A least absolute shrinkage and selection operator (LASSO) procedure selects basis functions for the projection. Tools are provided for simulation studies and empirical applications. The methods are based on Dong, Gao, Linton and Peng (2026) <doi:10.48550/arXiv.2607.05699>. |
| License: | MIT + file LICENSE |
| Encoding: | UTF-8 |
| URL: | https://github.com/Greatknee/SIVMethod |
| BugReports: | https://github.com/Greatknee/SIVMethod/issues |
| Depends: | R (≥ 3.5) |
| Imports: | glmnet, stats, utils |
| Suggests: | testthat (≥ 3.0.0) |
| Config/testthat/edition: | 3 |
| Config/roxygen2/version: | 8.0.0 |
| Date: | 2026-09-14 |
| NeedsCompilation: | no |
| Packaged: | 2026-09-15 11:48:25 UTC; greatknee |
| Author: | Yuhuai Chen [aut, cre], Chaohua Dong [aut], Jiti Gao [aut], Linton Oliver [aut], Bin Peng [aut] |
| Maintainer: | Yuhuai Chen <yuhuai.chen0351@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2026-09-27 16:00:09 UTC |
Semiparametric Instrumental Variable Estimation
Description
Estimation and inference for regression models with endogenous regressors using a semiparametric projection approach. Instrumental variables are constructed internally from observed regressors by projecting out a space of basis functions used to represent the conditional mean of the structural error. A least absolute shrinkage and selection operator (LASSO) procedure selects basis functions for the projection. See further details in Dong, Gao, Linton and Peng (2026) <https://arxiv.org/abs/2607.05699>.
Usage
SIV(Y, X, k)
Arguments
Y |
Response vector, n-by-1. |
X |
Regressor matrix, n-by-d. |
k |
Integer vector including the possible values of the truncation parameter, L_k-by-1. |
Value
A list containing:
- b_SIV_GCV, b_SIV_LSO
Regression coefficient estimates using the full basis and post-LASSO basis, respectively.
- gam_SIV_GCV, gam_LSO
Corresponding series coefficient estimates.
- sd_GCV, sd_LSO
Standard errors of the regression estimates.
- sd_gam
Standard errors of
gam_LSO.- e_GCV, e_LSO
Residuals from the respective fits.
- eps_LSO
Estimated
m(X_i) + e_i, assuming the first selected basis column is the constant term.- V_LSO
Post-LASSO basis matrix, including a constant column.
- Ind_LSO
Logical selection indicators for the original basis.
- m_all
Four-column matrix containing estimated
m(x), lower and upper pointwise 95% wild bootstrap confidence limits, and evaluation points, respectively.
Monte Carlo Simulation for Semiparametric Instrumental Variable Estimation
Description
Conducts Monte Carlo simulations for the designs in Example B.2.1. Compares ordinary least squares and the semiparametric projection method. Reports empirical bias, Monte Carlo standard deviations, and basis selection frequencies.
Usage
Sim(n, no_sim, case, k = NULL)
Arguments
n |
Integer. Number of observations. |
no_sim |
Integer. Number of simulation replications. |
case |
Character. One of "A", "B", "C" or "D". |
k |
Integer. The truncation parameter. |
Value
A list with:
- Output
Summary statistics matrix.
- Selection
Average selection frequencies.
Examples
Sim(100L, 200L, case = "A")
Monte Carlo Simulation for Testing Theory
Description
Evaluates the finite-sample size and power of the endogeneity test proposed in Section 3.2 using the simulation designs in Example B.2.3. Empirical rejection rates are reported across Monte Carlo replications.
Usage
Sim_test(n, no_sim, case = "A", c = 0.75, k = NULL)
Arguments
n |
Integer. Number of observations. |
no_sim |
Integer. Number of simulation replications. |
case |
Character. One of "A" or "B". |
c |
Numeric scalar with c >= 0. Used in
|
k |
Integer vector. The range of truncation parameters to be considered via Generalized Cross-Validation. |
Value
A list with:
- Output
Summary statistics matrix.
- Rejection_rate
Average test rejection rate.
Examples
Sim_test(100L, 200L, case = "A", c = 1)