bayesMCClust (version 1.0)

LMEntryPaperData:

Data From Fruehwirth-Schnatter et al. (2011): "Labor market entry and earnings dynamics: Bayesian inference using mixtures-of-experts Markov chain clustering"

Description

The empirical analysis in Fruehwirth-Schnatter et al. (2011) is based on data from the Austrian Social Security Database (ASSD), which combines detailed longitudinal information on employment and earnings of all private sector workers in Austria since 1972 (see References). The IEW Working Paper Zweimueller et al. (2009) (see Source) gives an overview and a description of the main characteristics of the Austrian Social Security Database.

The ASSD was made available for the Austrian Center of Labor Economics and the Analysis of the Welfare State (http://www.labornrn.at/). The considered sample consists of $N=49279$ male Austrian workers, who enter the labor market for the first time in the years 1975 to 1985 and are less than 25 years old at entry. The cohort analysis is based on an observation period from 1975 to 2005.

Usage

data(LMEntryPaperData)

Format

The format is:

List of 6
 $ InitValBetas: num [1:25, 1:4] 0 0 0 0 0 0 0 0 0 0 ...
  ..- attr(*, "dimnames")=List of 2
  .. ..$ : chr [1:25] "intercept" "unEmplRDist" "unskilled" "skilled" ...
  .. ..$ : chr [1:4] "h1" "h2" "h3" "h4"
 $ InitValClass: int [1:49279] 2 3 1 4 3 2 3 2 4 1 ...
 $ covariates  :'data.frame':   49279 obs. of  25 variables:
  ..$ intercept    : num [1:49279] 1 1 1 1 1 1 1 1 1 1 ...
  ..$ unEmplRDist  : num [1:49279] 0.91 0.697 0.905 0.91 1.051 ...
  ..$ unskilled    : num [1:49279] 0 0 0 0 0 0 0 1 0 0 ...
  ..$ skilled      : num [1:49279] 0 1 1 1 0 0 0 0 1 0 ...
  ..$ whiteColl    : num [1:49279] 0 0 1 0 1 0 0 1 1 1 ...
  ..$ wageCat1Dummy: num [1:49279] 1 1 1 0 0 1 1 1 0 0 ...
  ..$ wageCat2Dummy: num [1:49279] 0 0 0 0 1 0 0 0 0 1 ...
  ..$ wageCat3Dummy: num [1:49279] 0 0 0 1 0 0 0 0 1 0 ...
  ..$ wageCat4Dummy: num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ wageCat5Dummy: num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear76  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear77  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear78  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear79  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear80  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear81  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear82  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear83  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear84  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ entryYear85  : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ ia.ueRD.wc1D : num [1:49279] 0.91 0.697 0.905 0 0 ...
  ..$ ia.ueRD.wc2D : num [1:49279] 0 0 0 0 1.05 ...
  ..$ ia.ueRD.wc3D : num [1:49279] 0 0 0 0.91 0 ...
  ..$ ia.ueRD.wc4D : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
  ..$ ia.ueRD.wc5D : num [1:49279] 0 0 0 0 0 0 0 0 0 0 ...
 $ mccXiPrior  :List of 1
  ..$ :List of 1
  .. ..$ xi: num [1:6, 1:6] 0.7 0.15 0.0333 0.0333 0.0333 ...
 $ NjkiMat     : num [1:49279, 1:36] 0 0 0 2 7 0 4 0 0 1 ...
 $ Njk.i       : num [1:6, 1:6, 1:49279] 0 0 0 0 0 0 0 1 1 0 ...
  ..- attr(*, "dimnames")=List of 3
  .. ..$ : chr [1:6] "0" "1" "2" "3" ...
  .. ..$ : chr [1:6] "0" "1" "2" "3" ...
  .. ..$ : NULL
 

Source

The following IEW Working Paper gives an overview and a description of the main characteristics of the Austrian Social Security Database: Zweimueller, Josef, Winter-Ebmer, Rudolf, Lalive, Rafael, Kuhn, Andreas, Wuellrich, Jean-Philippe, Ruf, Oliver and Buechi, Simon, Austrian Social Security Database (May 4, 2009). Available at SSRN: http://ssrn.com/abstract=1399350 or at http://www.labornrn.at/wp/wp0903.pdf.

Details

LMEntryPaperData is a list containing the following objects:

InitValBetas

contains a matrix with the initial values (used in our paper) for the logit regression coefficients.

InitValClass

contains a vector with some initial values (used in our paper) for the classification variable (group membership for 4 groups).

covariates

contains the data.frame with the covariates used in the logit regression model. It contains the following variables:

		unEmplRDist
unemployment rate in the district
unskilled	dummy for unskilled workers
	skilled	dummy for skilled workers
		whiteColl
dummy for white collar workers
wageCat1Dummy,...,
	wageCat5Dummy	dummies for starting in the corresponding wage category
		entryYear76,...,
entryYear85	dummies for starting in the corresponding year
	ia.ueRD.wc1D,...,
		ia.ueRD.wc5D
interaction variable for unemployment rate in the
	district and the dummies for starting in the

mccXiPrior

contains the prior-parameters (used in the paper) for the transition matrices.

NjkiMat

contains the Njk.i-data in matrix format of dimension $49279x36$ (each row corresponds to the columns of the matrices in Njk.i).

Njk.i

contains the transition frequencies in a 3-dim array of dimension $6x6x49279$ containing the transition frequencies ($6 x 6$-matrices) of 49279 individuals. These represent the counts of transitions between wage categories from year to year with varying observation periods. Categories 1 to 5 correspond to the wage quintiles and 0 to no income.

References

Sylvia Fruehwirth-Schnatter, Christoph Pamminger, Andrea Weber and Rudolf Winter-Ebmer, (2011), "Labor market entry and earnings dynamics: Bayesian inference using mixtures-of-experts Markov chain clustering". Journal of Applied Econometrics. DOI: 10.1002/jae.1249 http://onlinelibrary.wiley.com/doi/10.1002/jae.1249/abstract

Link to Journal of Applied Econometrics Data Archive: http://econ.queensu.ca/jae/forthcoming/fruehwirth-schnatter-et-al/

See Also

mcClustExtended

Examples

data(LMEntryPaperData)
str(LMEntryPaperData)

# =====================   LMEntry Paper Data   =================================
#rm(list=ls(all=TRUE))

# set working directory
curDir <- getwd()

if ( !file.exists("bayesMCClust-wd") ) dir.create("bayesMCClust-wd")
setwd("bayesMCClust-wd") 
myOutfilesDir <- "LMEntry-Paper-Data-Outfiles"      
# ==============================================================================
if (!is.element("LMEntryPaperData$covariates", search())) { 
    attach(LMEntryPaperData$covariates)
}
# ==============================================================================
groupNr <- 4
# ==============================================================================
if ( FALSE ) {
  try(mcClustExtended(      # parameter lists (all four) must be complete!!!
     Data=list(dataFile=LMEntryPaperData$Njk.i, 
               storeDir=myOutfilesDir,
               priorFile= LMEntryPaperData$mccXiPrior,
               X = cbind( intercept=1, unEmplRDist, unskilled, skilled, whiteColl, 
                                    wageCat1Dummy, wageCat2Dummy, wageCat3Dummy, 
                                    wageCat4Dummy, wageCat5Dummy,
                                    entryYear76, entryYear77, entryYear78, 
                                    entryYear79, entryYear80, entryYear81, 
                                    entryYear82, entryYear83, entryYear84, 
                                    entryYear85,
                                    ia.ueRD.wc1D, ia.ueRD.wc2D, ia.ueRD.wc3D, 
                                    ia.ueRD.wc4D, ia.ueRD.wc5D
                        ) ),
     Prior=list(H=groupNr, 
                c=1,
                cOff=1,
                usePriorFile=TRUE,
                xiPooled=TRUE,
                N0=10,
                betaPrior = "informative", # N(0,1)
                betaPriorMean = 0,
                betaPriorVar = 1),
     Initial=list(xi.start.ind=3, 
                  pers=0.7,
                  S.i.start = LMEntryPaperData$InitValClass,
                  Beta.start = LMEntryPaperData$InitValBetas ), 
     Mcmc=list(M=15000,
               M0=10000,
               mOut=500,
               mSave=5000,
               seed=3546541) 
  ))
}

setwd(curDir)

if (is.element("LMEntryPaperData$covariates", search())) {
    detach(LMEntryPaperData$covariates)
}
# ==============================================================================