The Stata Journal: Promoting communications on statistics and Stata

We discuss the ivcrc command, which implements an instrumental-variables (IV) estimator for the linear correlated random-coefficients model. The correlated random-coefficients model is a natural generalization of the standard linear IV model that allows for endogenous, multivalued treatments and unobserved heterogeneity in treatment effects. The estimator implemented by ivcrc uses recent semiparametric identification results that allow for flexible functional forms and permit instruments that may be binary, discrete, or continuous. The ivcrc command also allows for the estimation of varying-coefficient regressions, which are closely related in structure to the proposed IV estimator. We illustrate the use of ivcrc by estimating the returns to education in the National Longitudinal Survey of Young Men.

In this article, I introduce new commands to calculate second-generation p-values (SGPVs) for common estimation commands in Stata. The sgpv command and its companions allow the easy calculation of SGPVs and their associated diagnostics, as well as the plotting of SGPVs against the standard p-values.

Eight multigroup segregation indices are decomposable into a between and a within term. They are two versions of 1) the mutual information index, 2) the symmetric Atkinson index, 3) the relative diversity index, and 4) Theil’s H index. In this article, we present the command dseg, which obtains all of them. It contributes to the stock of segregation commands in Stata by 1) implementing the decomposition in a single call, 2) providing the weights and local indices used in the computation of the within term, 3) facilitating the deployment of the decomposability properties of the eight indices in complex scenarios that demand tailor-made solutions, and 4) leveraging sample data with bootstrapping and approximate randomization tests. We analyze 2017 census data of public schools in the United States to illustrate the use of dseg. The subject topic is school racial segregation.

Ordinal responses can be generated, in a cross-sectional context, by different unobserved classes of population or, in a time-series context, by different latent regimes. We introduce a new command, swopit, that fits a mixture of ordered probit models with exogenous or endogenous switching between two latent classes (regimes). Switching is endogenous if unobservables in the classassignment model are correlated with unobservables in the outcome models. We provide a battery of postestimation commands; assess via Monte Carlo experiments the finite-sample performance of the maximum likelihood estimator of the parameters, probabilities, and their standard errors (both the asymptotic and bootstrap ones); and apply the new command to model the monetary policy interest rates.

Typical censoring models have mass points at the upper or lower tails, or at both tails, of an otherwise continuous outcome distribution. In contrast, we consider a censoring model with a mass point in the interior of the outcome distribution. We refer to this mass point as “bunching” and use it to estimate model parameters. For example, economic theory suggests that, for increasing marginal income tax rates, many taxpayers will report income exactly at the threshold where the tax rate increases. This translates into a censoring model with bunching at the threshold. The size of this mass point of taxpayers can be used to estimate an elasticity parameter that summarizes taxpayers’ responses to taxes. In this article, we introduce the command bunching, which implements new nonparametric and semiparametric identification methods for estimating elasticities developed by Bertanha, McCallum, and Seegert (2021, Technical Report 2021-002, Board of Governors of the Federal Reserve System). These methods rely on weaker assumptions than what are currently made in the literature and result in meaningfully different estimates of the elasticity.

In modern multivariate statistics, where high-dimensional datasets are ubiquitous, learning large (inverse-) covariance matrices is imperative for data analysis. A popular approach to estimating a large inverse-covariance matrix is to regularize the Gaussian log-likelihood function by imposing a convex penalty function. In a seminal article, Friedman, Hastie, and Tibshirani (2008, Biostatistics 9: 432–441) proposed a graphical lasso (Glasso) algorithm to efficiently estimate sparse inverse-covariance matrices from the convex regularized log-likelihood function. In this article, I first explore the Glasso algorithm and then introduce a new graphiclasso command for the large inverse-covariance matrix estimation. Moreover, I provide a useful command for tuning parameter selection in the Glasso algorithm using the extended Bayesian information criterion, the Akaike information criterion, and cross-validation. I demonstrate the use of Glasso using simulation results and real-world data analysis.

In this article, I introduce xtsfkk as a new command for fitting panel stochastic frontier models with endogeneity. The advantage of xtsfkk is that it can control for the endogenous variables in the frontier and the inefficiency term in a longitudinal setting. Hence, xtsfkk performs better than standard panel frontier estimators such as xtfrontier that overlook endogeneity by design. Moreover, xtsfkk uses Mata’s moptimize() functions for substantially faster execution and completion speeds. I also present a set of Monte Carlo simulations and examples demonstrating the performance and usage of xtsfkk.

In this article, we introduce a new community-contributed command called xtbunitroot, which implements the panel-data unit-root tests developed by Karavias and Tzavalis (2014, Computational Statistics and Data Analysis 76: 391–407). These tests allow for one or two structural breaks in deterministic components of the series and can be seen as panel-data counterparts of the tests by Zivot and Andrews (1992, Journal of Business and Economic Statistics 10: 251–270) and Lumsdaine and Papell (1997, Review of Economics and Statistics 79: 212–218). The dates of the breaks can be known or unknown. The tests allow for intercepts and linear trends, nonnormal errors, and cross-section heteroskedasticity and dependence. They have power against homogeneous and heterogeneous alternatives and can be applied to panels with small or large time-series dimensions.

In this article, we describe methodology that allows for multiple timescales using flexible parametric survival models without the need for time splitting. When one fits flexible parametric survival models on the log-hazard scale, numerical integration is required in the log likelihood to fit the model. The use of numerical integration allows incorporation of arbitrary functions of time into the model and hence lends itself to the inclusion of multiple timescales in an appealing way. We describe and exemplify these methods and show how to use the command stmt , which implements these methods, alongside its postestimation commands.

In this article, I show how to calculate consistent marginal effects on the original scale of the outcome variable in Stata after estimating a linear regression with a dependent variable that has been transformed by the inverse hyperbolic sine function. The method uses a nonparametric retransformation of the error term and accounts for any scaling of the dependent variable. The inverse hyperbolic sine function is not invariant to scaling, which is known to shift marginal effects between those from an untransformed dependent variable to those of a logtransformed dependent variable, when all observations are positive.

We introduce a new command, xtusreg, that estimates parameters of fixed-effects dynamic panel regression models under unequal time spacing. After reviewing the method, we examine the finite-sample performance of the command using simulated data. We also illustrate the command with the National Longitudinal Survey Original Cohorts: Older Men, whose personal interviews took place in the unequally spaced years of 1966, 1967, 1969, 1971, 1976, 1981, and 1990. The methods underlying xtusreg are those discussed by Sasaki and Xin (2017, Journal of Econometrics 196: 320–330).