Community-contributed Command Research Articles

Xtnumfac: A battery of estimators for the number of common factors in time series and panel-data models

In this article, we introduce a new community-contributed command, xtnumfac, for estimating the number of common factors in time-series and panel datasets using the methods of Bai and Ng (2002, Econometrica 70: 191–221), Ahn and Horenstein (2013, Econometrica 81: 1203–1227), Onatski (2010, Review of Economics and Statistics 92: 1004–1016), and Gagliardini, Ossola, and Scaillet (2019, Journal of Econometrics 212: 503–521). Common factors are usually unobserved or unobservable. In time series, they influence all predictors, while in paneldata models, they influence all cross-sectional units at different degrees. Examples are shocks from oil prices, inflation, or demand or supply shocks. Knowledge about the number of factors is key for multiple econometric estimation methods, such as Pesaran (2006, Econometrica 74: 967–1012), Bai (2009, Econometrica 77: 1229–1279), Norkute et al. (2021, Journal of Econometrics 220: 416–446), and Kripfganz and Sarafidis (2021, Stata Journal 21: 659–686). This article discusses a total of 10 methods to estimate the number of common factors. Examples based on Kapetanios, Pesaran, and Reese (2021, Journal of Econometrics 221: 510–541) show that U.S. house prices are exposed to up to 10 common factors. Therefore, when one fits models with house prices as a dependent variable, the number of factors must be considered.

Panel unit-root tests with structural breaks

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.

Travel distance and travel time using Stata: New features and major improvements in georoute

The community-contributed command georoute is designed to calculate travel distance and travel time between two addresses or two geographical points identified by their coordinates. Since its conception and description by Weber and Péclat (2017, Stata Journal 17: 962–971), the command has been gradually maintained and enriched. The new version of georoute presented in this article encompasses major improvements, such as the possibility to specify transport mode and departure time. The new features open the way to a multitude of more sophisticated research applications.

Effect sizes for contrasts of estimated marginal effects

The statistical literature is replete with calls to report standardized measures of effect size alongside traditional p-values and null hypothesis tests. While effect-size measures such as Cohen’s d and Hedges’s g are straightforward to calculate for t tests, this is not the case for parameters in more complex linear models, where traditional effect-size measures such as η 2 and ω 2 face limitations. After a review of effect sizes and their implementation in Stata, I introduce the community-contributed command mces. This postestimation command reports standardized effect-size statistics for dichotomous comparisons of marginal-effect contrasts obtained from margins and mimrgns, including with complex samples, for continuous outcome variables. mces provides Stata users the ability to report straightforward estimates of effect size in many modeling applications.

Implementing quantile selection models in Stata

In this article, we describe qregsel, a community-contributed command that implements a copula-based sample-selection correction for quantile regression recently proposed by Arellano and Bonhomme (2017, Econometrica 85: 1–28). The command allows the user to model selection in quantile regressions by using either a Gaussian or a one-dimensional Frank copula. We illustrate the use of qregsel with two examples. First, we apply the method to the fictional dataset used in the Stata Base Reference Manual for the heckman command. Second, we replicate part of the empirical application of the original article using data for the United Kingdom that cover the period 1978–2000 to compare wages of males and females at different quantiles.

Meeting assumptions in the estimation of reliability

Researchers and psychometricians have long used Cronbach’s α as a measure of reliability. However, there have been growing calls to replace Cronbach’s α with measures that have more defensible assumptions. One of the most common and straightforward recommended reliability estimates is ω. After a review of reliability and its estimation in Stata, I introduce the community-contributed command omegacoef. This command reports McDonald’s ω in a format similar to the base alpha command. omegacoef provides Stata users the ability to easily compute estimates of reliability with the confidence that the necessary statistical assumptions are met.

Stacked linear regression analysis to facilitate testing of hypotheses across OLS regressions

In empirical work, researchers frequently test hypotheses of parallel form in several regressions, which raises concerns about multiple testing. One way to address the multiple-testing issue is to jointly test the hypotheses (for example, Pei, Pischke, and Schwandt [2019, Journal of Business & Economic Statistics 37: 205–216] and Lee and Lemieux [2010, Journal of Economic Literature 48: 281–355]). While the existing commands suest (Weesie, 1999, Stata Technical Bulletin Reprints 9: 231–248) and mvreg enable Stata users to follow this approach, both are limited in several dimensions. For instance, mvreg assumes homoskedasticity and uncorrelatedness across sampling units, and neither command is designed to be used with panel data. In this article, we introduce the new community-contributed command stackreg, which overcomes the aforementioned limitations and allows for some settings and features that go beyond the capabilities of the existing commands. To achieve this, stackreg runs an ordinary least-squares regression in which the regression equations are stacked as described, for instance, in Wooldridge (2010, Econometric Analysis of Cross Section and Panel Data, p. 166–173, MIT Press) and applies cluster–robust variance–covariance estimation.

Testing for slope heterogeneity in Stata

In this article, we introduce a new community-contributed command, xthst, to test for slope heterogeneity in panels with many observations over cross-sectional units and time periods. The command implements such a test, the delta test (Pesaran and Yamagata, 2008, Journal of Econometrics 142: 50–93). Under its null, slope coefficients are homogeneous across cross-sectional units. Under the alternative, slope coefficients are heterogeneous in the cross-sectional dimension. xthst also includes two extensions. The first is a heteroskedasticity- and autocorrelation-consistent robust test along the lines of Blomquist and Westerlund (2013, Economics Letters 121: 374–378). The second extension is a cross-sectional-dependence robust version. We discuss all tests and present examples using an economic growth model. A Monte Carlo simulation shows that the size and the power behave as expected.

Nonparametric synthetic control using the npsynth command

In this article, I build on the work of Abadie and Gardeazabal (2003, American Economic Review 93: 113–132) and Abadie, Diamond, and Hainmueller (2010, Journal of the American Statistical Association 105: 493–505), extending the synthetic control method for program evaluation—implemented in Stata via the community-contributed command synth—to the case of a nonparametric identification of the synthetic (or counterfactual) time pattern of a treated unit (a country, a region, a city, etc.) subject to a specific intervention in a given time. After theoretical description of the model, I present npsynth, the command I developed for estimating the nonparametric synthetic control method proposed in this article. Using both simulated and real data, I set out a comparison of the performance of the parametric and nonparametric methods and widely discuss the results.

Feologit: A new command for fitting fixed-effects ordered logit models

In this article, we describe how to fit panel-data ordered logit models with fixed effects using the new community-contributed command feologit. Fixed-effects models are increasingly popular for estimating causal effects in the social sciences because they flexibly control for unobserved time-invariant heterogeneity. The ordered logit model is the standard model for ordered dependent variables, and this command is the first in Stata specifically for this model with fixed effects. The command includes a choice between two estimators, the blowup and cluster (BUC) estimator introduced in Baetschmann, Staub, and Winkelmann (2015, Journal of the Royal Statistical Society, Series A 178: 685–703) and the BUC- τ estimator in Baetschmann (2012, Economics Letters 115: 416–418). Baetschmann, Staub, and Winkelmann (2015) showed that the BUC estimator has good properties and is almost as efficient as more complex estimators such as generalized method-of-moments and empirical likelihood estimators. The command and model interpretations are illustrated with an analysis of the effect of parenthood on life satisfaction using data from the German Socio-Economic Panel.

Estimation of pre- and posttreatment average treatment effects with binary time-varying treatment using Stata

In this article, we describe tvdiff, a community-contributed command that implements a generalization of the difference-in-differences estimator to the case of binary time-varying treatment with pre- and postintervention periods. tvdiff is flexible and can accommodate many actual situations, enabling the user to specify the number of pre- and postintervention periods and a graphical representation of the estimated coefficients. In addition, tvdiff provides two distinct tests for the necessary condition of the identification of causal effects, namely, two tests for the so-called parallel-trend assumption. tvdiff is intended to simplify applied works on program evaluation and causal inference when longitudinal data are available.

Intcount: A command for fitting count-data models from interval data

In this article, I describe a community-contributed command, intcount, that fits one of several regression models for count data observed in interval form. The models available are Poisson, negative binomial, and binomial, and they can be fit in standard or zero-inflated form. I illustrate the command with an application to analysis of data from the UK Understanding Society survey on the demand for healthcare services.

Mixmcm: A community-contributed command for fitting mixtures of Markov chain models using maximum likelihood and the EM algorithm

Markov chain models and finite mixture models have been widely applied in various strands of the academic literature. Several studies analyzing dynamic processes have combined both modeling approaches to account for unobserved heterogeneity within a population. In this article, we describe mixmcm, a community-contributed command that fits the general class of mixed Markov chain models, accounting for the possibility of both entries into and exits from the population. To account for the possibility of incomplete information within the data (that is, unobserved heterogeneity), the model is fit with maximum likelihood using the expectation-maximization algorithm. mixmcm enables users to fit the mixed Markov chain models parametrically or semiparametrically, depending on the specifications chosen for the transition probabilities and the mixing distribution. mixmcm also allows for endogenous identification of the optimal number of homogeneous chains, that is, unobserved types or “components”. We illustrate mixmcm‘s usefulness through three examples analyzing farm dynamics using an unbalanced panel of commercial French farms.

Bootstrap pointwise confidence intervals for covariate-adjusted survivor functions in the Cox model

Survival functions are a common visualization of predictions from the Cox model. However, neither Stata’s stcurve command nor the communitycontributed scurve tvc command allows one to estimate confidence intervals. In this article, I discuss how bootstrap confidence intervals can be formed for covariate-adjusted survival functions in the Cox model. The new bsurvci command automates this procedure and allows users to visualize the results. bsurvci enables one to estimate uncertainty around survival functions estimated from Cox models with time-varying coefficients, a capability that was not previously available in Stata. Furthermore, it provides Stata users with an additional option for survival estimates from Cox models with proportional hazards by allowing them to choose between bootstrap confidence intervals using bsurvci and asymptotic confidence intervals from an existing community-contributed command, survci. Because asymptotic confidence intervals make distributional assumptions when constructing confidence intervals, the bootstrap procedure proposed in this article provides a nonparametric alternative.

Candle charts for financial technical analysis

Technical analysis is an important part of financial industry, research, and teaching. The methodology has two parts: i) calculation of the individual tools and ii) visual representations. In this article, I provide a community-contributed command, candlechart, to draw the most common technical analysis charts. My intent is to draw these charts similarly to industry examples. The popular candle price chart is combined with charts for volume, moving-average convergence divergence, relative strength index, and Bollinger bands.

Content Analysis: Frequency Distribution of Words

Many academic fields use content analysis. At the core of most common content analysis lies frequency distribution of individual words. Websites and documents are mined for usage and frequency of certain words. In this article, we introduce a community-contributed command, wordfreq, to process content (online and local) and to prepare a frequency distribution of individual words. Additionally, another community-contributed command, wordcloud, is introduced to draw a simple word cloud graph for visual analysis of the frequent usage of specific words.

Inference for Clustered Data

In this article, we introduce clusteff, a community-contributed command for checking the severity of cluster heterogeneity in cluster–robust analyses. Cluster heterogeneity can cause a size distortion leading to underrejection of the null hypothesis. Carter, Schnepel, and Steigerwald (2017, Review of Economics and Statistics 99: 698–709) develop the effective number of clusters to reflect a reduction in the degrees of freedom, thereby mirroring the distortion caused by assuming homogeneous clusters. clusteff generates the effective number of clusters. We provide a decision tree for cluster–robust analysis, demonstrate the use of clusteff, and recommend methods to minimize the size distortion.

Event Study Estimations Using Stata: The Estudy Command

In this article, we introduce the community-contributed command estudy and illustrate how it can be used to perform an event study customizing the statistical framework, from the estimates of abnormal returns to the tests for their statistical significance. Our command significantly improves the existing commands in terms of both completeness and user comprehension.

Baselinetable: A Command for Creating one- and Two-way Tables of Summary Statistics

In this article, I describe baselinetable, a community-contributed command for creating one- and two-way tables of summary statistics for a list of numeric variables. Unlike other Stata tabulation commands such as tabulate, table, or tabstat, the variables forming the table rows can be a mixture of continuous (summarized by mean, standard deviation, etc.) and categorical (summarized by percentages and frequencies). baselinetable provides considerable flexibility in the way the results are displayed. In particular, the summary statistics and their presentation can be different for each row variable. Additional features include several options for displaying counts of missing and nonmissing data points and the ability to restrict results to subgroups of the data for individual row variables. The contents of the table can be saved as a data file or text file, or they can be exported to Excel. The motivation for baselinetable is the descriptive table commonly seen in health research publications in which the baseline characteristics of two or more groups are compared. This descriptive table usually has only one column for each group, generally with at least two summary statistics in each column (for example, mean and standard deviation for continuous variables or frequency and percentage for categorical variables). The baselinetable command supports reproducible research by enabling researchers to easily create tables whose contents can be used unchanged in publications.

Fitting and Interpreting Correlated Random-coefficient Models Using Stata

In this article, we introduce the community-contributed command randcoef, which fits the correlated random-effects and correlated random-coefficient models discussed in Suri (2011, Econometrica 79: 159–209). While this approach has been around for a decade, its use has been limited by the computationally intensive nature of the estimation procedure that relies on the optimal minimum distance estimator. randcoef can accommodate up to five rounds of panel data and offers several options, including alternative weight matrices for estimation and inclusion of additional endogenous regressors. We also present postestimation analysis using sample data to facilitate understanding and interpretation of results.