Package 'feisr' reference manual

Title:	Estimating Fixed Effects Individual Slope Models
Description:	Provides the function feis() to estimate fixed effects individual slope (FEIS) models. The FEIS model constitutes a more general version of the often-used fixed effects (FE) panel model, as implemented in the package 'plm' by Croissant and Millo (2008) <doi:10.18637/jss.v027.i02>. In FEIS models, data are not only person demeaned like in conventional FE models, but detrended by the predicted individual slope of each person or group. Estimation is performed by applying least squares lm() to the transformed data. For more details on FEIS models see Bruederl and Ludwig (2015, ISBN:1446252442); Frees (2001) <doi:10.2307/3316008>; Polachek and Kim (1994) <doi:10.1016/0304-4076(94)90075-2>; Ruettenauer and Ludwig (2020) <doi:10.1177/0049124120926211>; Wooldridge (2010, ISBN:0262294354). To test consistency of conventional FE and random effects estimators against heterogeneous slopes, the package also provides the functions feistest() for an artificial regression test and bsfeistest() for a bootstrapped version of the Hausman test.
Authors:	Tobias Ruettenauer [aut, cre] , Volker Ludwig [aut]
Maintainer:	Tobias Ruettenauer <[email protected]>
License:	GPL (>= 2)
Version:	1.3.0
Built:	2024-10-31 18:37:23 UTC
Source:	https://github.com/ruettenauer/feisr

Estimating Fixed Effects Individual Slope Models

Description

The main purpose of the package feisr is the estimation of fixed effects individual slopes models and respective test statistics. The fixed effects individual slopes (FEIS) estimator is a more general version of the well-known fixed effects estimator (FE), which allows to control for heterogeneous slopes in addition to time-constant heterogeneity (Bruederl and Ludwig 2015; Ruettenauer and Ludwig 2020; Wooldridge 2010). This is done by running an lm() model on pre-transformed data, where we (1) estimate the individual-specific predicted values for the dependent variable and each covariate based on an individual intercept and the additional slope variables, (2) detrend the original data by these individual-specific predicted values, and (3) run an OLS model on the residual data. The package also provides two specification test for heterogeneous slopes (more details and examples can be found in Ruettenauer and Ludwig 2020).

Details

The main functions of the feisr package are:

- feis(): fixed effects individual slopes estimator by applying lm to detrended data.

- feistest(): regression-based Hausman test for fixed effects individual slope models.

- bsfeistest(): bootstrapped Hausman test for fixed effects individual slope models.

The functions included in the R package feisr are also available in the xtfeis ado (https://ideas.repec.org/c/boc/bocode/s458045.html) for Stata. The plm-package provides functions for estimation of related models, like the mean group (MG) or common correlated effects mean groups (CCEMG) estimator via pmg or models with variable coefficients via pvcm.

Author(s)

Tobias Ruettenauer

Volker Ludwig

References

Bruederl J, Ludwig V (2015). “Fixed-Effects Panel Regression.” In Best H, Wolf C (eds.), The Sage Handbook of Regression Analysis and Causal Inference, 327–357. Sage, Los Angeles. ISBN 1446252442.

Ruettenauer T, Ludwig V (2020). “Fixed Effects Individual Slopes: Accounting and Testing for Heterogeneous Effects in Panel Data or Other Multilevel Models.” Sociological Methods and Research, OnlineFirst. ISSN 0049-1241, doi:10.1177/0049124120926211.

Wooldridge JM (2010). Econometric Analysis of Cross Section and Panel Data. MIT Press, Cambridge, Mass. ISBN 0262294354.

Bootstrapped Regression Test

Description

Estimates a bootstrapped Hausman test for fixed effects individual slope models.

Usage

bsfeistest(
  model = NA,
  type = c("all", "bs1", "bs2", "bs3"),
  terms = NULL,
  rep = 500,
  seed = NULL,
  prog = TRUE,
  ...
)
bsfeistest(
  model = NA,
  type = c("all", "bs1", "bs2", "bs3"),
  terms = NULL,
  rep = 500,
  seed = NULL,
  prog = TRUE,
  ...
)

Arguments

`model`	an object of class "`feis`".
`type`	one of "`all`" (the Default), "`bs1`" for test of FEIS against FE only, "`bs2`" for test of FE against RE only, and "`bs3`" for test of FEIS against RE only (see also Details).
`terms`	An optional character vector specifying which coefficients should be jointly tested. By default, all covariates are included in the Wchi-squared test. For "`type=art2`", the slope variable is always included in "`terms`".
`rep`	the number of repetitions to be used in bootstrapping (default is 500).
`seed`	the seed used for random sampling in bootstrapping. Needs to be a valid integer. If not specified, the current seed is used.
`prog`	... logical. If `TRUE` (the Default) shows the progress in the output window.
`...`	further arguments.

Details

The function computes a bootstrapped version of the Hausman test (Hausman 1978). Pairs cluster bootstrapping (Cameron et al. 2008; Ruettenauer and Ludwig 2020) is used to obtain the empirical variance-covariance matrix of the estimators, either for FEIS and conventional FE, convention FE and RE, or FEIS and RE.

type="bs1" estimates a bootstrapped Hausman test comparing fixed effects individual slope models and conventional fixed effects models. In this case, bsfeistest tests for inconsistency of the convetional FE model due to heterogeneous slopes. type="bs2" estimates a bootstrapped version of the well-known Hausman test comparing conventional fixed effects models against random effects models. type="bs3" estimates a bootstrapped Hausman directly comparing FEIS against RE, thereby testing for inconsistency of the RE model due to either heterogeneous slopes or time-constant omitted heterogeneity. Bootstrapping is perfomed by resampling with replacement while keeping the number of groups identical to the number of groups in the original dataset. A wald test from aod package is used to perform a Wald chi-squared test on the differences between coefficients.

Value

An object of class "feistest", containing the following elements:

`wald_feis`	an object of class "`wald.test`" testing the fixed effects individual slopes model against the conventional fixed effects model (`type="bs1"`).
`wald_fe`	an object of class "`wald.test`" testing the fixed effects model against the random effects model (`type="bs2"`).
`wald_re`	an object of class "`wald.test`" testing the fixed effects individual slopes model against the random effects model (`type="bs3"`).
`vcov1`	the empirical (bootstrapped) variance-covariance matrix of the coefficients obtained from FEIS and FE (`type="bs1"`).
`vcov2`	the empirical (bootstrapped) variance-covariance matrix of the coefficients obtained from FE and RE (`type="bs2"`).
`vcov3`	the empirical (bootstrapped) variance-covariance matrix of the coefficients obtained from FEIS and RE (`type="bs3"`).
`bscoef.feis`	a matrix containing the estimated FEIS coefficients of each bootstrap run.
`bscoef.fe`	a matrix containing the estimated FE coefficients of each bootstrap run.
`bscoef.re`	a matrix containing the estimated RE coefficients of each bootstrap run.
`call`	the matched call.
`formula`	an object of class "`Formula`" describing the model.
`type`	the type of performed test(s).
`sample`	a list containing the IDs sampled in each run.
`seed`	the seed used for bootstrapping.
`terms`	character vector of covariates are included in the Wchi-squared test.

References

Cameron AC, Gelbach JB, Miller DL (2008). “Bootstrap-Based Improvements for Inference with Clustered Errors.” Review of Economics and Statistics, 90(3), 414–427. ISSN 0034-6535, doi:10.1162/rest.90.3.414.

Hausman JA (1978). “Specification Tests in Econometrics.” Econometrica, 46(6), 1251–1271. ISSN 00129682.

Ruettenauer T, Ludwig V (2020). “Fixed Effects Individual Slopes: Accounting and Testing for Heterogeneous Effects in Panel Data or Other Multilevel Models.” Sociological Methods and Research, OnlineFirst. ISSN 0049-1241, doi:10.1177/0049124120926211.

Examples

data("mwp", package = "feisr")
## Not run: 
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id", robust = TRUE)
bsht <- bsfeistest(feis.mod, type = "bs1", rep = 100, seed = 1234)
summary(bsht)

## End(Not run)

data("mwp", package = "feisr")
## Not run: 
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id", robust = TRUE)
bsht <- bsfeistest(feis.mod, type = "bs1", rep = 100, seed = 1234)
summary(bsht)

## End(Not run)

Detrend data by individual slopes

Description

Detrends the input data by the predicted values based on the slope parameters within each group specified by id. The result is equal to the transformed data used for estimation in feis.

Usage

detrend(
  data,
  slopes,
  id = NULL,
  intercept = TRUE,
  na.action = c("na.exlude", "na.omit"),
  tol = .Machine$double.eps,
  predicted = FALSE,
  ...
)
detrend(
  data,
  slopes,
  id = NULL,
  intercept = TRUE,
  na.action = c("na.exlude", "na.omit"),
  tol = .Machine$double.eps,
  predicted = FALSE,
  ...
)

Arguments

`data`	a `data.frame`, `matrix`, or `vector` of data to be detrended. If `id` and / or `slopes` are given as character (see below), must contain `id` and / or `slopes` as variable(s). Otherwise must be excluded.
`slopes`	a `data.frame`, `matrix`, or `vector` of slopes to be used for detrending, not containing an intercept. Optionally, a `character` vector of the names of slope variables in `data`. For pure de-meaning use `"slopes = 1"`.
`id`	a `vector` of a unique group / person identifier. Optionally, a `character` of the name of the unique group / person identifier in `data`. For overall detrending, use `"id = 1"`.
`intercept`	logical. If `TRUE` the slopes will contain an individual intercept (default is `TRUE`). For `"id = 1"`, this is an overall intercept. Ignored if "slopes = 1".
`na.action`	character, either `na.exclude` (default) or `na.omit` indicates the use of `NA`s. `na.exclude` passes `NA`s through to the output (same length as input). `na.omit` drops `NA` rows (list-wise).
`tol`	the tolerance for detecting linear dependencies in the residual maker transformation (see `solve`).
`predicted`	logical. If `TRUE` returns the predicted values instead of the detrended data (default is `FALSE`).
`...`	further arguments.

Details

detrend performs within-group "residual maker" transformation on the origin data. Within each group, the predicted values of the columns in data are computed based on the slope columns plus an individual intercept if intercept = TRUE (the default). Subsequently the predicted values are subtracted from the origin data. The transformed data can, for instance, be used to obtain coefficients of a fixed effects individual slopes estimator via lm (Bruederl and Ludwig 2015; Ruettenauer and Ludwig 2020; Wooldridge 2010).

Estimation requires at least q+1 observations per unit, where q is the number of slope parameters (including a constant). detrend automatically selects only those groups from the current data set which have at least q+1 observations, and returns NA for all groups with n_i < q+1.

NA values in the input data are handled by list-wise deletion based on the data to be detrended and the slopes.

Value

An object of class "data.frame" or "numeric (if only one data column), containing the detrended data with row.names equal to the row.names of the origin data. If input is an unnamed vector, names are 1:length.

References

Examples

data("mwp", package = "feisr")

# Detrend entire data.frame
mwp_det <- detrend(data = mwp, slopes = c("exp", "expq"), id = "id")

# Detrend single variable
lnw_det <- detrend(data = mwp$lnw, slopes = mwp[, c("exp", "expq")], id = mwp$id)

data("mwp", package = "feisr")

# Detrend entire data.frame
mwp_det <- detrend(data = mwp, slopes = c("exp", "expq"), id = "id")

# Detrend single variable
lnw_det <- detrend(data = mwp$lnw, slopes = mwp[, c("exp", "expq")], id = mwp$id)

Fixed Effects Individual Slope Estimator

Description

Estimates fixed effects individual slope estimators by applying linear lm models to "detrended" data.

Usage

feis(
  formula,
  data,
  id,
  weights = NULL,
  robust = FALSE,
  intercept = FALSE,
  dropgroups = FALSE,
  tol = .Machine$double.eps,
  ...
)

## S3 method for class 'feis'
formula(x, lhs = NULL, rhs = NULL, ...)

## S3 method for class 'feis'
terms(x, lhs = NULL, rhs = NULL, ...)

## S3 method for class 'feis'
residuals(object, ...)

## S3 method for class 'feis'
df.residual(object, ...)

## S3 method for class 'feis'
coef(object, ...)

## S3 method for class 'feis'
sigma(object, ...)

## S3 method for class 'feis'
deviance(object, ...)

## S3 method for class 'feis'
nobs(object, ...)

## S3 method for class 'feis'
fitted(object, ...)

## S3 method for class 'feis'
hatvalues(model, ...)
feis(
  formula,
  data,
  id,
  weights = NULL,
  robust = FALSE,
  intercept = FALSE,
  dropgroups = FALSE,
  tol = .Machine$double.eps,
  ...
)

## S3 method for class 'feis'
formula(x, lhs = NULL, rhs = NULL, ...)

## S3 method for class 'feis'
terms(x, lhs = NULL, rhs = NULL, ...)

## S3 method for class 'feis'
residuals(object, ...)

## S3 method for class 'feis'
df.residual(object, ...)

## S3 method for class 'feis'
coef(object, ...)

## S3 method for class 'feis'
sigma(object, ...)

## S3 method for class 'feis'
deviance(object, ...)

## S3 method for class 'feis'
nobs(object, ...)

## S3 method for class 'feis'
fitted(object, ...)

## S3 method for class 'feis'
hatvalues(model, ...)

Arguments

`formula`	a symbolic description for the model to be fitted (see Details).
`data`	a `data.frame` containing the specified variables.
`id`	the name of a unique group / person identifier (as string).
`weights`	an optional vector of weights to be used in the fitting process. See `lm`.
`robust`	logical. If `TRUE` estimates cluster robust standard errors (default is `FALSE`).
`intercept`	logical. If `TRUE` estimates the model with an intercept (default is `FALSE`).
`dropgroups`	logical. If `TRUE` groups without any within variance on a slope variable are dropped , if `FALSE` those variables are omitted for the respective groups only (default is `FALSE`).
`tol`	the tolerance for detecting linear dependencies in the residual maker transformation (see `solve`). The argument is forwarded to `bsfeistest`.
`...`	further arguments.
`lhs`, `rhs`	indexes of the left- and right-hand side for the methods formula and terms.
`object`, `x`, `model`	an object of class "`feis`".

Details

feis is a special function to estimate linear fixed effects models with individual-specific slopes. In contrast to conventional fixed effects models, data are not person "demeaned", but "detrended" by the predicted individual slope of each person (Bruederl and Ludwig 2015; Ruettenauer and Ludwig 2020; Wooldridge 2010).

Estimation requires at least q+1 observations per unit, where q is the number of slope parameters (including a constant). feis automatically selects only those groups from the current data set which have at least q+1 observations. The function returns a warning if units with <q+1 observations are dropped.

The function requires a two-part formula, in which the second part indicates the slope parameter(s). If, for example, the model is y ~ x1 + x2, with the slope variables x3 and x4, the model can be estimated with:

formula = y ~ x1 + x2 | x3 + x4

To estimate a conventional fixed effects model without individual slopes, please use y ~ x1 + x2 | 1 to indicate that the slopes should only contain an individual-specific intercept.

If specified, feis estimates panel-robust standard errors. Panel-robust standard errors are robust to arbitrary forms of serial correlation within groups formed by id as well as heteroscedasticity across groups (see Wooldridge 2010, pp. 379-381).

The model output can be exported using the texreg package.

Value

An object of class "feis", containing the following elements:

`coefficients`	the vector of coefficients.
`vcov`	the scaled (if specified, robust) variance-covariance matrix of the coefficients. See `vcov.feis` for unscaled vcov

`residuals`	the vector of residuals (computed from the "detrended" data).
`df.residual`	degrees of freedom of the residuals.
`formula`	an object of class "`Formula`" describing the model.
`model`	the original model frame as a `data.frame` containing the original variables used for estimation.
`modelhat`	a constructed model frame as a `data.frame` containing the predicted values from the first stage regression using the slope variable(s) as predictor(s).
`modeltrans`	a constructed model frame as a `data.frame` containing the "detrended" variables used for the final model estimation. Note that the weights are already used for detrending if specified.
`response`	the vector of the "detrended" response variable.
`fitted.values`	the vector of fitted values (computed from the "detrended" data).
`id`	a vector containing the unique person identifier.
`weights`	a vector containing weights used in fitting, or integer 1 if not speficied in call.
`call`	the matched call.
`assign`	assign attributes of the formula.
`na.omit`	(where relevant) a vector of the omitted observations. The only handling method of `NA`s is "`omit`".
`contrasts`	(only where relevant) the contrasts used.
`arg`	a list containing the used methods. Only "`feis`" and "`individual`" effects available.
`slopevars`	a character vector containing the names of the slope variables.
`r2`	R squared of the "detrended" model.
`adj.r2`	adjusted R squared of the "detrended" model.
`vcov_arg`	a character containing the method used to compute the variance-covariance matrix.
`tol`	the tolerance parameter (for use in bsfeistest).

References

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol + as.factor(yeargr) | exp + I(exp^2),
                 data = mwp, id = "id", robust = TRUE)
summary(feis.mod)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol + as.factor(yeargr) | exp + I(exp^2),
                 data = mwp, id = "id", robust = TRUE)
summary(feis.mod)

Artificial Regression Test

Description

Estimates a regression-based Hausman test for fixed effects individual slope models.

Usage

feistest(
  model = NA,
  robust = FALSE,
  type = c("all", "art1", "art2", "art3"),
  terms = NULL,
  ...
)
feistest(
  model = NA,
  robust = FALSE,
  type = c("all", "art1", "art2", "art3"),
  terms = NULL,
  ...
)

Arguments

`model`	an object of class "`feis`".
`robust`	logical. If `TRUE` uses cluster robust standard errors (Default is `FALSE`).
`type`	one of "`all`" (the Default), "`art1`" for test of FEIS against FE only, "`art2`" for test of FE against RE only, and "`art3`" for test of FEIS against RE only (see also Details).
`terms`	An optional character vector specifying which coefficients should be jointly tested. By default, all covariates are included in the Wchi-squared test. For "`type=art2`", the slope variable is always included in "`terms`".
`...`	further arguments.

Details

The Hausman test can be computed by estimating a correlated random effects model (see Wooldridge 2010, pp. 328-334, Ruettenauer and Ludwig 2020). This is achieved by estimating a Mundlak (Mundlak 1978) specification using random effects models with plm. Subsequently, feistest tests whether the time-constant variables / slope variables are correlated with the unobserved heterogeneity by using a Wald chi-squared test.

type="art1" estimates an extended regression-based Hausman test comparing fixed effects individual slope models and conventional fixed effects models. For art1 the Mundlak-specification (Mundlak 1978) includes the person-specific averages, but additionally the person-specific slope estimates used for "detrending" in feis. This allows to test whether we can omit the estimated values based on the slopes and reduce the model to a conventional FE model. The Wald test of type="art1" is applied to the slope variables only. type="art2" estimates the conventional regression-based Hausman test (as described in Wooldridge 2010, pp. 328-334) comparing conventional fixed effects models against random effects models. type="art3" estimates a regression-based Hausman test comparing FEIS directly against RE, thereby testing for inconsistency of the RE model due to either heterogeneous slopes or time-constant omitted heterogeneity. For art3 the Mundlak-specification includes only the person-specific slopes, and no averages. This allows to test whether we can omit the estimated values based on the slopes and reduce the model to a conventional RE model. (for a formal description please see Ruettenauer and Ludwig 2020).

Currently, the tol option in feis() is only forwarded in bsfeistest, but not in feistest.

If specified (robust=TRUE), feistest uses panel-robust standard errors.

Value

An object of class "feistest", containing the following elements:

`wald_feis`	an object of class "`wald.test`" testing the fixed effects individual slopes model against the conventional fixed effects model (`type="art1"`).
`wald_fe`	an object of class "`wald.test`" testing the fixed effects model against the random effects model (`type="art2"`).
`wald_re`	an object of class "`wald.test`" testing the fixed effects individual slopes model against the random effects model (`type="art3"`).
`vcov1`	the variance-covariance matrix of CREIS (`type="art1"`).
`vcov2`	the variance-covariance matrix of CRE (`type="art2"`).
`vcov3`	the variance-covariance matrix of CREIS without the means (`type="art3"`).
`CREIS`	an object of class "`plm`" (see `plm`) estimating a Correlated Random Effect Individual Slope model (`type="art1"`).
`CRE`	an object of class "`plm`" (see `plm`) estimating a Correlated Random Effect model (`type="art2"`).
`CREIS2`	an object of class "`plm`" (see `plm`) estimating a Correlated Random Effect Individual Slope model without including the covariates' means (`type="art3"`).
`call`	the matched call.
`robust`	logical. If `TRUE` cluster robust standard errors were used (Default is `FALSE`.
`formula`	an object of class "`Formula`" describing the model.
`type`	the type of performed test(s).
`terms`	character vector of covariates are included in the Wchi-squared test.

References

Mundlak Y (1978). “On the Pooling of Time Series and Cross Section Data.” Econometrica, 46(1), 69. ISSN 00129682.

Ruettenauer T, Ludwig V (2020). “Fixed Effects Individual Slopes: Accounting and Testing for Heterogeneous Effects in Panel Data or Other Multilevel Models.” Sociological Methods and Research, OnlineFirst. ISSN 0049-1241, doi:10.1177/0049124120926211.

Wooldridge JM (2010). Econometric Analysis of Cross Section and Panel Data. MIT Press, Cambridge, Mass. ISBN 0262294354.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id", robust = TRUE)
ht <- feistest(feis.mod, robust = TRUE, type = "all")
summary(ht)
# Only for marry coefficient
ht2 <- feistest(feis.mod, robust = TRUE, type = "all", terms = c("marry"))
summary(ht2)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id", robust = TRUE)
ht <- feistest(feis.mod, robust = TRUE, type = "all")
summary(ht)
# Only for marry coefficient
ht2 <- feistest(feis.mod, robust = TRUE, type = "all", terms = c("marry"))
summary(ht2)

model.matrix for feis objects

Description

Methods to extract transformed model matrix for "feis" objects.

Usage

## S3 method for class 'feis'
model.matrix(object, ...)
## S3 method for class 'feis'
model.matrix(object, ...)

Arguments

`object`	an object of class "`feis`".
`...`	further arguments.

Details

model.matrix for feis objects returns the model or design matrix of the respective FEIS model. This is the transformed (de-trended) data, which is used for estimation of the model in lm().

Value

An object of class "matrix" for model.matrix.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + as.factor(yeargr) | exp,
                 data = mwp, id = "id")
mm <- model.matrix(feis.mod)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + as.factor(yeargr) | exp,
                 data = mwp, id = "id")
mm <- model.matrix(feis.mod)

A function to extract the model.response

Description

Returns the de-trended response variable of a feis object.

Usage

model.response.feis(x, ...)
model.response.feis(x, ...)

Arguments

`x`	an object of class `feis`.
`...`	further arguments.

Details

The function provides a convenient way to return the model.response of a feis object. This is the transformed (de-trended) variable which is used for estimation of the final model.

Value

A "numeric" of the transformed response variable of the estimation model.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id")
y_tilde <- model.response.feis(feis.mod)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | year,
                 data = mwp, id = "id")
y_tilde <- model.response.feis(feis.mod)

Panel data including wages and family status

Description

A random sample from the National Longitudinal Survey of Youth (Bureau of Labor Statistics 2014). It contains information on wages, family status, and work experience for a random sample of men. For a description of the original dataset and variable construction see Ludwig and Bruederl (2018).

Usage

mwp
mwp

Format

A data frame with 3100 observations and 17 variables:

id: unique person identifier
year: survey year
lnw: natural log of hourly wage rate
exp: work experience in current job, in years
expq: work experience in current job squared
marry: family status (=0 if not married, =1 if married)
evermarry: indicator if ever married (=0 if never married, =1 if married at some point)
enrol: current erolment in education (=0 not enroled, =1 enroled)
yeduc: years of formal education
age: respondents current age
cohort: respondents birth cohort
yeargr: grouped year (1=1979-1980, 2=1981-1985, 3=1986-1990, 4=1991-1995, 5=1996-2000)
yeargr1: dummy indicating grouped year=1
yeargr2: dummy indicating grouped year=2
yeargr3: dummy indicating grouped year=3
yeargr4: dummy indicating grouped year=4
yeargr5: dummy indicating grouped year=5

Source

Ludwig and Bruederl (2018)

References

Bureau of Labor Statistics (2014). National Longitudinal Survey of Youth 1979 Cohort, 1979-2012 (rounds 1-23). Center for Human Resource Research, The Ohio State University, Columbus, OH.

Ludwig V, Bruederl J (2018). “Is There a Male Marital Wage Premium? New Evidence from the United States.” American Sociological Review, 83(4), 744–770. ISSN 0003-1224, doi:10.1177/0003122418784909.

Predict method for feis models

Description

Predicted values based on linear model object.

Usage

## S3 method for class 'feis'
predict(
  object,
  newdata = NULL,
  se.fit = FALSE,
  vcov = NULL,
  interval = c("none", "confidence", "prediction"),
  level = 0.95,
  pred.var = sigma_sq,
  ...
)
## S3 method for class 'feis'
predict(
  object,
  newdata = NULL,
  se.fit = FALSE,
  vcov = NULL,
  interval = c("none", "confidence", "prediction"),
  level = 0.95,
  pred.var = sigma_sq,
  ...
)

Arguments

`object`	an object of class "`feis`", fitted model.
`newdata`	an optional data frame in which to look for variables with which to predict. If omitted, the fitted values are used.
`se.fit`	a switch indicating if standard errors are required.
`vcov`	optional variance-covariance matrix for std.err. calculation.
`interval`	type of interval calculation.
`level`	tolerance/confidence level.
`pred.var`	the variance for future observations to be assumed for prediction intervals. By default, equals the residual variance
`...`	further arguments.

Details

predict.lm produces predicted values, obtained by evaluating the regression function in the frame newdata (which defaults to model.matrix(object)). If the logical se.fit is TRUE, standard errors of the predictions are calculated. If the vcov is not provided, the object$vcov is used, thus allowing for robust variance-covariance matrices. Setting intervals specifies computation of confidence or prediction (tolerance) intervals at the specified level.

Note: Currently, predictions are based on the transformed (de-trended) data.

Value

A vector of predictions or a matrix of predictions and bounds with column names fit, lwr, and upr if interval is set.

Examples

feis.mod <- feis(lnw ~ age | exp,
data = mwp, id = "id", robust = TRUE)

new <- data.frame(age = seq(-10, 10, 1))
feis.pred <- predict(feis.mod, newdata = new,
                     se.fit = TRUE, interval = "confidence")

## Not run: 
matplot(new$age, feis.pred$fit, lty = c(1,2,2),
        type = "l", ylab = "predicted y")

## End(Not run)

feis.mod <- feis(lnw ~ age | exp,
data = mwp, id = "id", robust = TRUE)

new <- data.frame(age = seq(-10, 10, 1))
feis.pred <- predict(feis.mod, newdata = new,
                     se.fit = TRUE, interval = "confidence")

## Not run: 
matplot(new$age, feis.pred$fit, lty = c(1,2,2),
        type = "l", ylab = "predicted y")

## End(Not run)

Extract individual slopes

Description

Extracts the individual slopes (alpha_i) from a feis object created by feis.

Usage

slopes(model = NA, ...)
slopes(model = NA, ...)

Arguments

`model`	an object of class "`feis`".
`...`	further arguments.

Details

The function extracts a matrix containing the individual slope parameters (alpha_i), which equals the coefficient(s) of regressing the depenent variable on the slope parameter(s).

If slope variables are perfectly collinear within a cluster, one variable is dropped and the function returns 0 for the respective slope and cluster.

Value

An N x J matrix containing the individual slopes for each cluster unit N and slope variable J. Rownames indicate the cluster id.

Examples

data("Produc", package = "plm")
feis.mod <- feis("log(gsp) ~ log(pcap) + log(pc) + log(emp) + unemp | year",
                 data = Produc, id = "state", robust = TRUE)
slps <- slopes(feis.mod)

data("Produc", package = "plm")
feis.mod <- feis("log(gsp) ~ log(pcap) + log(pc) + log(emp) + unemp | year",
                 data = Produc, id = "state", robust = TRUE)
slps <- slopes(feis.mod)

Summary for feis objects

Description

The summary method for feis objects generates some additional information about estimated feis models.

Usage

## S3 method for class 'feis'
summary(object, vcov = NULL, ...)

## S3 method for class 'summary.feis'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  subset = NULL,
  ...
)
## S3 method for class 'feis'
summary(object, vcov = NULL, ...)

## S3 method for class 'summary.feis'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  subset = NULL,
  ...
)

Arguments

`object`	an object of class "`feis`".
`vcov`	a variance-covariance matrix furnished by the user or a function to calculate one.
`...`	further arguments.
`x`	an object of class "`summary.feis`".
`digits`	number of digits for printed output.
`width`	the maximum length of the lines in the printed output.
`subset`	a character or numeric vector indicating a subset of the table of coefficients to be printed.

Value

An object of class "summary.feis", containing the elements of the feis object (see feis). The object is forwarded to print method. The following objects are modified:

`coefficients`	a matrix with the estimated coefficients, standard errors, t-values, and p-values, if argument vcov is NULL the standard errors are calculated by the `vcov` in the input object.
`r.squared`	a vector containing R squared and adjusted R squared.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry | exp,
                 data = mwp, id = "id")
summary(feis.mod)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry | exp,
                 data = mwp, id = "id")
summary(feis.mod)

Summary for feistest and bsfeistest objects

Description

The summary method for feistest and bsfeistest objects prints the results of Artificial Regression Tests or Bootstrapped Hausman Tests for FEIS models.

Usage

## S3 method for class 'feistest'
summary(object, ...)

## S3 method for class 'summary.feistest'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  ...
)

## S3 method for class 'bsfeistest'
summary(object, ...)

## S3 method for class 'summary.bsfeistest'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  ...
)
## S3 method for class 'feistest'
summary(object, ...)

## S3 method for class 'summary.feistest'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  ...
)

## S3 method for class 'bsfeistest'
summary(object, ...)

## S3 method for class 'summary.bsfeistest'
print(
  x,
  digits = max(3, getOption("digits") - 2),
  width = getOption("width"),
  ...
)

Arguments

`object`	an object of class "`feistest`" or "`bsfeistest`".
`...`	further arguments.
`x`	an object of class "`summary.feistest`" or "`summary.bsfeistest`".
`digits`	number of digits for printed output.
`width`	the maximum length of the lines in the printed output.

Value

An object of class "summary.feistest" or "summary.bsfeistest", equal to the original input object (see feistest and bsfeistest). The object is forwarded to print method.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry | exp,
                 data = mwp, id = "id")
ht <- feistest(feis.mod, robust = TRUE, type = "all")
summary(ht)

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry | exp,
                 data = mwp, id = "id")
ht <- feistest(feis.mod, robust = TRUE, type = "all")
summary(ht)

Calculate Variance-Covariance Matrix for feis models

Description

Returns the variance-covariance matrix of the main parameters of an object of class "feis". By default, this is the unscaled variance-covariance matrix.

Usage

## S3 method for class 'feis'
vcov(object, ..., scale = FALSE)
## S3 method for class 'feis'
vcov(object, ..., scale = FALSE)

Arguments

`object`	an object of class "`feis`", fitted model.
`...`	further arguments.
`scale`	logical. If `TRUE` returns scaled vcov by sigma^2 (default is `FALSE`).

Details

By default, vcov() return the unscaled variance-covariance matrix of the fitted FEIS model. If set to scale = TRUE, the vcov is scaled by the nuisance parameter sigma^2 (as is object$vcov). Note that corrections for clustering (i.e. robust = TRUE in the fitted model) are ignored in vcov(). In this case, object$vcov will return the vcov with corrections for clustering.

Value

A matrix of the estimated covariances between the parameter estimates in the fitted FEIS model.

Examples

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | exp,
                 data = mwp, id = "id")
vcov(feis.mod)
all.equal(vcov(feis.mod), feis.mod$vcov) # FALSE: not equal, because vcov() unscaled
all.equal(vcov(feis.mod, scale = TRUE), feis.mod$vcov) # equal

data("mwp", package = "feisr")
feis.mod <- feis(lnw ~ marry + enrol | exp,
                 data = mwp, id = "id")
vcov(feis.mod)
all.equal(vcov(feis.mod), feis.mod$vcov) # FALSE: not equal, because vcov() unscaled
all.equal(vcov(feis.mod, scale = TRUE), feis.mod$vcov) # equal

Package 'feisr'

Help Index

Estimating Fixed Effects Individual Slope Models

Description

Details

Author(s)

References

See Also

Bootstrapped Regression Test

Description

Usage

Arguments

Details

Value

References

See Also

Examples

Detrend data by individual slopes

Description

Usage

Arguments

Details

Value

References

See Also

Examples

Fixed Effects Individual Slope Estimator

Description

Usage

Arguments

Details

Value

References

See Also

Examples

Artificial Regression Test

Description

Usage

Arguments

Details

Value

References

See Also

Examples

model.matrix for feis objects

Description

Usage

Arguments

Details

Value

See Also

Examples

A function to extract the model.response

Description

Usage

Arguments

Details

Value

Examples

Panel data including wages and family status

Description

Usage

Format

Source

References

Predict method for feis models

Description

Usage

Arguments

Details

Value

See Also

Examples

Extract individual slopes

Description

Usage

Arguments

Details

Value

Examples