xtreg2way: Estimation of Two Way Fixed Effects Model

In xtreg2way: Feasible Estimation of the Two-Way Fixed Effect Model

Description

xtreg2way estimates a 2-way fixed effect model absorbing the two set of dummies and reports standard errors.

Usage

xtreg2way(y, ...)

## S3 method for class 'formula'
xtreg2way(
  formula,
  data,
  iid = NULL,
  tid = NULL,
  w = NULL,
  struc = NULL,
  se = "",
  noise = "",
  ...
)

## Default S3 method:
xtreg2way(
  y,
  X,
  iid = NULL,
  tid = NULL,
  w = NULL,
  struc = NULL,
  se = "",
  noise = "",
  ...
)

Arguments

y	Either a numeric of data, or a formula
...	Other parameters, based on method used
formula	Model specifications
data	A dataframe with labels from the formula y
iid	(optional) The group ID, size N-by-1 - not needed if struc is provided
tid	(optional) The time ID, size N-by-1 - not needed if struc is provided
w	(optional) The vector of weights, size N-by-1. If omitted, w will be set to 1 for all observations
struc	(optional) This list contains the results from the first step of the algorithm. To save computational time, you can rerun the algorithm on different columns by providing this struc.
se	(optional) This indicates standard error estimate to be calculcated. Possible values include: se=="0" : standard errors assuming homoscedasticity and no within group correlation or serial correlation. se=="1" : standard errors proposed by Arellano (1987) robust to heteroscedasticity and serial correlation. se=="2" : standard errors robust to heteroscedasticity but assumes no correlation within group or serial correlation. se=="11" : Arellano standard errors with a degree of freedom correction performed by Stata xtreg, fe. If se is omitted or set to [] then it is set to 1 and the Arellano (1987) estimator is computed.
noise	(optional) If noise is set to "1", then results are displayed
X	A matrix of data

Value

betaHat (K-by-1) vector of estimated coefficients

aVarHat (K-by-K) estimate of the matrix of variances and covariance of the estimator.

y (N-by-1) the residual of the projection of y on the two sets of dummies.

X (N-by-K) the residual of the projection of each column of X on the two sets of dummies.

struc (list) results of the first step of the algorithm.

Methods (by class)

• formula: This function ingests a formula as the first argument, and requires data as a data.frame

• default: Default Method

Examples

hhid <- c("a","b","c","a","b","c" ,"a","b","c" ,"a","b","c" ,"a","b","c")
tid <- c("1","1" ,"1" ,"2","2" ,"3","3","3" ,"4","4","5" ,"5","6","6" ,"6")
w <- rep(1, 15)
x1 <- rnorm(15, mean=50, sd=10)   
x2 <- rnorm(15, mean=50, sd=10)   
y <- x1 + rnorm(15, mean=50, sd=10)    
#The most basic way to use this function
output <- xtreg2way(y, x1, hhid, tid, w, se="2", noise="1")
#You can rerun faster with different columns using output$struc
output2 <- xtreg2way(y, data.frame(x1,x2), struc=output$struc)
#Or you can use a formula and specify data=
output3 <- xtreg2way(y~x1+x2, data=data.frame(x1,x2,y), iid=hhid, tid=tid, w=w, 
                     se="2", noise="1")
           

xtreg2way documentation built on Aug. 11, 2020, 9:06 a.m.