Abstract: prodest is a new and comprehensive Stata module for production function estimation based on the control function approach. It includes Olley-Pakes (OP), Levinshon-Petrin (LP), Wooldridge (WRDG) and Ackerberg-Caves-Frazer (ACF) estimation techniques, plus a brand new methodology (Mollisi-Rovigatti, MR) in order to better deal with short panels. Its basic usage is similar to that of existing modules like opreg or levpet, but adds many features to control the optimization procedures and address estimation issues - gross output vs. value added, endogenous variables, attrition in the data.

Abstract: kappaetc calculates various measures of interrater agreement along with their standard errors and confidence intervals. Statistics are calculated for any number of raters, any number of categories, and in the presence of missing values (i.e. varying number of raters per subject). Disagreement among raters may be weighted by user-defined weights or a set of prerecorded weights, suitable for any level of measurement.

Abstract: aextlogit is a wrapper for xtlogit which estimates the fixed effects logit and reports estimates of the average (semi-) elasticities of Pr(y=1|x,u) with respect to the regressors, and the corresponding standard errors and t-statistics. The method used to compute the (semi-) elasticities was first described by Kitazawa (2012, "Hyperbolic transformation and average elasticity in the framework of the fixed effects logit model," Theoretical Economics Letters, 2, 192-199.).

Abstract: qregpd can be used to fit the quantile regression for panel data (QRPD) estimator developed in Powell (2015). The estimator addresses a fundamental problem posed by alternative fixed-effect quantile estimators: inclusion of individual fixed effects alters the interpretation of the estimated coefficient on the treatment variable. As detailed in Powell(2016), the QRPD estimator is a special case of the generalized quantile estimator implemented by genqreg. Numerical optimization proceeds via a Nelder-Mead algorithm. As estimation and calculation of standard errors can sometimes pose numerical challenges, the user can estimate generalized quantile regressions using Markov Chain Monte Carlo methods or grid-search methods.

Abstract: xtcd2 tests residuals or a variable for weak cross sectional dependence in a panel data model. The test is based on the test outlined in Pesaran (Econometric Reviews, 2015). Under the null the residuals or the variable are weakly cross sectional dependent. The package supports balanced and unbalanced panels and can be performed as a postestimation command.

Abstract: marhis generates predictive margins and marginal effects plots with a histogram summarizing the distribution of the variable on the x-axis.

Abstract: ppml_panel_sg is a "fast" Poisson Pseudo-maximum Likelihood estimation command for use with international data and other types of spatial flows. It is specifically designed to alleviate the computational burden of the many fixed effects required by structural gravity models, particularly the many "pair" fixed effects that are required in order to consistently estimate the effects of trade policies in panel settings. Key features include a check to verify the existence of estimates, the allowance for pair-specific time trends, and the ability to store fixed effects post-estimation for use in structural work.

Abstract: absdid estimates a treatment effect using the semiparametric difference in difference estimator presented by Alberto Abadie in his seminal paper Abadie, A. (2005), "Semiparametric Difference-in-Differences Estimators", Review of Economic Studies 72, 1-19.

Abstract: While Stata's computational capabilities have intensively increased over the last decade, the quality of its default figure schemes is still a matter of debate amongst users. Clearly some of the arguments speaking against Stata figures are subject to individual taste, but others are not, such as for instance: horizontal labelling, unnecessary background tinting, missing gridlines, oversized markers. The two schemes introduced here attempt to solve the major shortcomings of Stata's default figure schemes. The schemes come with 21 new colors, of which seven colors are distinguishable for people suffering from color blindness. This package provides users with four new figure schemes: plotplain (plain and simple plotting environment, avoids chartjunk); plotplainblind (plain and simple plotting environment, avoids chartjunk + colorblind friendly); plottig (replicates R ggplot in most regards); plottigblind (replicates R ggplot in most regards + colorblind friendly)

Abstract: stddiff calculates the standardized difference between two groups for both continuous and categorical variables. Standardized difference estimates are increasingly used to describe to compare groups in clinical trials and observational studies, in preference over p-values. While Stata has some commands to calculate standardized differences for continuous variables, it does not currently have a corresponding command for categorical variables.

Abstract: xtdcce2 estimates a heterogeneous coefficient model in a dynamic panel with dependence between cross sectional units. It supports the Common Correlated Effects Estimator (CCE) by Pesaran (2006), the Dynamic Common Correlated Effects Estimator (DCCE), proposed by Chudik and Pesaran (2015) and the Mean Group Estimator (MG, Pesaran and Smith, 1995) and the Pooled Mean Group Estimator (PMG, Shin et. Al 1999). In addition, the estimation of long run coefficients using the Cross-Sectional Augmented Distributed Lag (CS-DL, Chudik et al. 2016) and Cross-Sectional Augmented ARDL (CS-ARDL, Chudik et al. 2016) are implemented. xtdcce2 tests for cross sectional dependence (if xtcd2 installed), estimates the exponent of cross-sectional dependence (if xtcse2 installed) following Bailey, Kapetanios and Pesaran (2016) and supports instrumental variable estimations (if ivreg2 installed).

Abstract: Panel data make it possible both to control for unobserved confounders and to include lagged, endogenous regressors. Trying to do both at the same time, however, leads to serious estimation difficulties. In the econometric literature, these problems have been solved by using lagged instrumental variables together with the generalized method of moments (GMM). In Stata, commands such as xtabond and xtdpdsys have been used for these models. xtdpdml addresses the same problems via maximum likelihood estimation implemented with Stata's structural equation modeling (sem) command. The ML (sem) method is substantially more efficient than the GMM method when the normality assumption is met and suffers less from finite sample biases. xtdpdml greatly simplifies the SEM model specification process; makes it possible to test and relax many of the constraints that are typically embodied in dynamic panel models; unlike most related methods, allows for the inclusion of time-invariant variables in the model; takes advantage of Stata's ability to use full information maximum likelihood (FIML) for dealing with missing data; provides an overall goodness of fit measure by default and provides easy access to others; and can also generate code for use with Mplus or lavaan.

Abstract: genqreg can be used to fit the generalized quantile regression estimator developed in Powell (2016). The generalized quantile estimator addresses a fundamental problem posed by traditional quantile estimators: inclusion of additional covariates alters the interpretation of the estimated coefficient on the treatment variable. As detailed in Powell (2016), the generalized quantile estimator implemented by genqreg addresses this problem and produces unconditional quantile treatment effects even in the presence of additional control variables. Numerical optimization proceeds via a Nelder-Mead algorithm. As estimation and calculation of standard errors can sometimes pose numerical challenges, the user can estimate generalized quantile regressions using Markov Chain Monte Carlo methods or grid-search methods.

Abstract: pstrata stratifies the propensity score into an optimal number of quantiles (meaning, strictly, quantile-based bins), where optimal means that no statistical differences are found on the propensity score between treatment groups within any quantile. pstrata generates strata variables for binary as well as multiple treatments.

Abstract: ftools consists of a Mata file and several Stata commands: The Mata file creates identifiers (factors) from variables by using hash functions instead of sorting the data, so it runs in time O(N) and not in O(N log N). The Stata commands exploit this to avoid sort operations, at the cost of being slower for small datasets (mainly because of the cost involved in moving data from Stata to Mata). Implemented commands are fcollapse, fegen group, and fsort. Note that most of the capabilities of levels and contract are already supported by these commands. Possible commands include more egen functions and merge and reshape alternatives.

Abstract: silhouette calculates and graphs the silhouette width for the cluster solution given by the grouping variable, using the pairwise distance matrix given in the distmat option. Silhouette width is an indicator of cluster adequacy. It compares for each case, the mean distance to other cases in the cluster in which the case is, and the mean distance to the nearest neighbour cluster. Silhouette widths less than zero indicate a case that fits poorly in its cluster.

Abstract: xtqptest calculates the bias-corrected Q(P) statistic for serial correlation described in Born & Breitung (Econometric Reviews, 2016) for varlist of ue-residuals.

Abstract: iati downloads and converts aid activity data from the International Aid Transparency Initiative (IATI) into Stata format. Users can select from various aid donor and implementing organizations who publish their data to IATI and import aid activity information directly into Stata for analysis.

Abstract: georoute calculates the georouting distance between two addresses or two geographical points identified by their coordinates. It uses the HERE API (see https://developer.here.com) to retrieve distances in two steps. In the first step, addresses are geocoded and their geographical coordinates (latitude and longitude) are obtained. In the second step, the georouting distance between the two points is obtained. The user can also directly provide geographical coordinates, which will bypass the first step.

Abstract: igenerate generates new indicator variables from categorical predictors, including weighted effect coded interactions. Note that Stata’s built-in command contrast does this job for most coding schemes and I recommend using contrast for most situations in applied research. igenerate does, however, construct weighted effect coded interactions, something contrast cannot do. For more information on this approach see the paper “Weighted effect coded interaction effects: a novel moderated regression model for observational studies” by Te Grotenhuis et al. /2016) in the International Journal of Public Health. Furthermore, igenerate might be useful for didactical purposes.

Abstract: cospectdens estimates several quantities commonly used in bivariate spectral analysis such as cospectrum, coherency-squared, phase spectrum and gain spectrum. varlist accepts two time series: the first variable is treated as the output variable (y), and the second variable is treated as the input variable (x). Smoothing is directly applied to individual periodogram and cross-periodogram obtained from the FFT of variables. Endpoints are adjusted cyclically in the central moving average smoothing. Users may supply their weights as an option or choose one of the weighting schemes. Current version does not apply tapering. This command only graphs the coherency-squared which may be interpreted as the frequency-domain counterpart of the correlation coefficient. To graph other measures users may request them to be saved in an output file.

Abstract: regxfe estimates a linear high order fixed effect, allowing for up to 7 fixed effects. It allows for the use of weights, robust and one way clustered standard errors. Robust and cluster errors are estimated based on the same assumptions as in the regress and areg commands. The command is based on a pre-transformation of the variables in order to absorb the effect of the FE variables of all dependent and independent variables before the model is estimated. The transformed dataset can be saved as a separate file. The degrees of freedom to estimate the standard errors are corrected using an approximation of the number of non-identifiable parameters, applying a modified Abowd, Creecy, and Kramarz (2002) algorithm.

Abstract: wtdttt estimates parameters of the ordinary and reverse Waiting Time Distribution (WTD) by maximum likelihood. Covariates can be used to improve precision of predicted prescription durations. Diagnostic plots can be obtained to assess fit of the model.

Abstract: sdmxuse imports data from statistical agencies using the SDMX standard. Available providers are the European Central Bank (ECB), Eurostat (ESTAT), the International Monetary Fund (IMF), the Organisation for Economic Co-operation and Development (OECD), the United Nations Statistics Division (UNSD) and the World Bank (WB). You can get a complete list of publicly available datasets from a provider by specifying the resource: dataflow. Then, you can obtain the Data Structure Definition (DSD) of a given dataset by specifying the resource: datastructure. Finally, you can download the dataset by specifying the resource: data.

Abstract: bdiff perform several tests (Fisher's Permutation test; Seemingly Unrelated Regression test, see suest) to determine the significance of observed differences in coefficient estimates between two groups. By default, bdiff performs traditional Fisher's permutation test (sampling without replacement) of differences in coefficient estimates between two groups It uses simulation evidence to determine the significance of observed differences in coefficient estimates between two groups. For general introduction of this method, see Efron and Tibshirani (1993, Section 15.2, pp.202).