Wooldridge Source: F. Vella (1993), “A Simple Estimator for Simultaneous Models with Censored Endogenous Regressors,” International Economic Review 34, 441-457. Professor Vella kindly provided the data. Data loads lazily.
data('fringe')A data.frame with 616 observations on 39 variables:
annearn: annual earnings, $
hrearn: hourly earnings, $
exper: years work experience
age: age in years
depends: number of dependents
married: =1 if married
tenure: years with current employer
educ: years schooling
nrtheast: =1 if live in northeast
nrthcen: =1 if live in north central
south: =1 if live in south
male: =1 if male
white: =1 if white
union: =1 if union member
office:
annhrs: annual hours worked
ind1: industry dummy
ind2:
ind3:
ind4:
ind5:
ind6:
ind7:
ind8:
ind9:
vacdays: $ value of vac. days
sicklve: $ value of sick leave
insur: $ value of employee insur
pension: $ value of employee pension
annbens: vacdays+sicklve+insur+pension
hrbens: hourly benefits, $
annhrssq: annhrs^2
beratio: annbens/annearn
lannhrs: log(annhrs)
tenuresq: tenure^2
expersq: exper^2
lannearn: log(annearn)
peratio: pension/annearn
vserat: (vacdays+sicklve)/annearn
https://www.cengage.com/cgi-wadsworth/course_products_wp.pl?fid=M20b&product_isbn_issn=9781111531041
Currently, this data set is used in only one Computer Exercise – to illustrate the Tobit model. It can be used much earlier. First, one could just ignore the pileup at zero and use a linear model where any of the hourly benefit measures is the dependent variable. Another possibility is to use this data set for a problem set in Chapter 4, after students have read Example 4.10. That example, which uses teacher salary/benefit data at the school level, finds the expected tradeoff, although it appears to less than one-to-one. By contrast, if you do a similar analysis with FRINGE.RAW, you will not find a tradeoff. A positive coefficient on the benefit/salary ratio is not too surprising because we probably cannot control for enough factors, especially when looking across different occupations. The Michigan school-level data is more aggregated than one would like, but it does restrict attention to a more homogeneous group: high school teachers in Michigan.
Used in Text: page 624-625
str(fringe)
#> 'data.frame': 616 obs. of 39 variables:
#> $ annearn : num 15000 6500 6909 5512 7800 ...
#> $ hrearn : num 7.81 5 2.35 4.5 3.75 ...
#> $ exper : int 14 7 22 2 19 24 42 21 22 40 ...
#> $ age : int 36 23 38 18 35 40 58 37 37 59 ...
#> $ depends : int 2 0 3 0 0 3 0 1 5 0 ...
#> $ married : int 1 0 1 0 0 1 1 1 0 1 ...
#> $ tenure : num 15 8 0.5 0.5 2 8 8 0.5 15 8 ...
#> $ educ : int 18 10 6 12 12 13 12 13 10 6 ...
#> $ nrtheast: int 1 0 0 0 0 1 1 0 1 0 ...
#> $ nrthcen : int 0 0 0 0 0 0 0 1 0 0 ...
#> $ south : int 0 1 1 1 1 0 0 0 0 1 ...
#> $ male : int 1 1 1 1 1 1 1 1 1 1 ...
#> $ white : int 1 1 1 1 0 1 1 1 1 1 ...
#> $ union : int 0 0 0 0 0 0 1 0 1 1 ...
#> $ office : int 1 0 0 0 0 0 0 0 0 0 ...
#> $ annhrs : num 1920 1300 2940 1225 2080 ...
#> $ ind1 : int 1 0 0 0 0 0 0 0 0 0 ...
#> $ ind2 : int 0 1 1 1 1 1 1 1 1 1 ...
#> $ ind3 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind4 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind5 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind6 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind7 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind8 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ ind9 : int 0 0 0 0 0 0 0 0 0 0 ...
#> $ vacdays : num 975 0 0 0 0 ...
#> $ sicklve : num 421 0 0 0 0 ...
#> $ insur : num 618 0 0 0 0 ...
#> $ pension : num 1368 0 0 0 0 ...
#> $ annbens : num 3381 0 0 0 0 ...
#> $ hrbens : num 1.76 0 0 0 0 ...
#> $ annhrssq: num 3686400 1690000 8643600 1500625 4326400 ...
#> $ beratio : num 0.225 0 0 0 0 ...
#> $ lannhrs : num 7.56 7.17 7.99 7.11 7.64 ...
#> $ tenuresq: num 225 64 0.25 0.25 4 64 64 0.25 225 64 ...
#> $ expersq : int 196 49 484 4 361 576 1764 441 484 1600 ...
#> $ lannearn: num 9.62 8.78 8.84 8.61 8.96 ...
#> $ peratio : num 0.0912 0 0 0 0 ...
#> $ vserat : num 0.0931 0 0 0 0 ...
#> - attr(*, "time.stamp")= chr "25 Jun 2011 23:03"