Skip to contents

Repeated cross-sections of workers from the May Current Population Survey (CPS) for 1974, 1975, 1977 and 1978, used by Gruber (1994) to study who pays for state laws that required health insurance to cover maternity. It is a classic triple-differences (DDD) application: mandate states versus comparison states, before versus after the laws, and married women aged 20 to 40 versus a comparison group of workers.

Usage

gruber1994

Format

A data frame with 27,653 rows and 29 columns:

id

Integer. Row identifier, unique for each row. Each row is a different worker.

year

Integer. Year of the May CPS: 1974 or 1975 (before the mandates) and 1977 or 1978 (after).

state

Character. State: IL, NJ or NY (mandate states) or OH, IN, CT, MA or NC (comparison states).

treatment_starts

Integer. First treated period: 1976 for IL, NJ and NY, whose mandates took effect between July 1976 and January 1977; 0 for the comparison states (never treated).

log_wage

Numeric. Log of hourly_wage.

log_hours

Numeric. Log of hours_per_week.

hourly_wage

Numeric. Hourly wage in 1978 dollars (see Details).

hours_per_week

Numeric. Hours worked per week.

education

Numeric. Years of education completed: the highest grade attended minus 1, minus 1 more if that grade was not completed, and 0 if negative.

experience

Numeric. Potential experience in years: age minus education minus 6, set to 0 if negative.

nonwhite

Integer. 1 if the worker is not white, 0 otherwise.

union

Integer. 1 if the worker is a union member, 0 otherwise.

treat

Integer. Same as treat_group.

exp

Integer. Same as exp_state.

post

Integer. Same as post_period.

treat_group

Integer. 1 if the worker is a married woman aged 20 to 40 (spouse present or armed-forces spouse present), the pregnancy-relevant group; 0 for the comparison group (workers older than 40 and never-married men aged 20 to 40).

exp_state

Integer. 1 if the state passed a maternity mandate (IL, NJ, NY), 0 otherwise.

post_period

Integer. 1 for 1977 and 1978 (after the mandates), 0 for 1974 and 1975.

industry_3

Integer. 1 if the worker is in construction, 0 otherwise.

industry_4

Integer. 1 if the worker is in manufacturing, 0 otherwise.

industry_5

Integer. 1 if the worker is in transportation, 0 otherwise.

industry_6

Integer. 1 if the worker is in wholesale trade, 0 otherwise.

industry_7

Integer. 1 if the worker is in retail trade, 0 otherwise.

industry_8

Integer. 1 if the worker is in finance, 0 otherwise.

industry_9

Integer. 1 if the worker is in business services, 0 otherwise.

industry_10

Integer. 1 if the worker is in personal services, 0 otherwise.

industry_11

Integer. 1 if the worker is in entertainment, 0 otherwise.

industry_12

Integer. 1 if the worker is in professional services, 0 otherwise.

industry_13

Integer. 1 if the worker is in public administration, 0 otherwise.

Source

NBER extracts of the May Current Population Survey for 1974, 1975, 1977 and 1978. Prepared by the package authors with the sample restrictions of Gruber (1994).

Details

Illinois, New Jersey and New York passed laws that required health insurance to cover maternity like other conditions. The laws took effect between July 1976 and January 1977. Gruber (1994) compares the change in wages of married women aged 20 to 40 with the change for a comparison group of workers, in the mandate states and in comparison states (OH and IN for IL; CT, MA and NC for NJ and NY).

The sample follows Gruber (1994, footnote 15): workers aged 20 to 65, not self-employed, with an hourly wage of at least $1 and at most $100 in 1978 dollars, and with non-missing education, experience, race, union status and industry; only the treatment and comparison groups; only the eight states above.

For workers paid by the hour, hourly_wage is the reported hourly rate; for other workers it is weekly earnings divided by weekly hours. Nominal wages are converted to 1978 dollars with the consumer price index (49.3 in 1974, 53.8 in 1975, 60.6 in 1977 and 65.2 in 1978).

The industry columns group 1970 Census industry codes into major industries. Agriculture and mining are the omitted groups: a worker with all industry columns equal to 0 is in one of these two industries.

References

Gruber, J. (1994). The Incidence of Mandated Maternity Benefits. American Economic Review, 84(3), 622-641.

Examples

# DR-DDD with repeated cross-sections (one row per worker)
out <- ddd(yname = "log_wage", tname = "year", idname = "id",
           gname = "treatment_starts", pname = "treat_group",
           xformla = ~ nonwhite + union + education + experience,
           data = gruber1994, control_group = "nevertreated",
           base_period = "universal", est_method = "dr", panel = FALSE)
agg_ddd(out, type = "simple")
#>  Call:
#> agg_ddd(ddd_obj = out, type = "simple")
#> ========================= DDD Aggregation ============================
#>  Overall ATT:
#>    ATT    Std. Error     [ 95%  Conf. Int.] 
#>  0.004        0.0387    -0.0717      0.0798 
#> 
#>  Note: * indicates that the confidence interval does not contain zero.
#>  --------------------------- Data Info   -----------------------------
#>  Outcome variable: log_wage
#>  Qualification variable: treat_group
#>  Control group:  Never Treated
#>  --------------------------- Std. Errors  ----------------------------
#>  Level of significance:  0.05
#>  Analytical standard errors.
#>  =====================================================================
#>  See Ortiz-Villavicencio and Sant'Anna (2025) for details.