David Graziano, Xiao Liang, Cory Mohling, Andrew Reid, Ken Sulma, Elizabeth Welch 30 years of gender pay gap consistency Between 1979 and 1989 female-to-male pay ratio rose 9% Convergence slowed in 1990’s rising only 3.5% Dataset: Michigan Panel Study of Income Dynamics (PSID) ◦ Only dataset that contains information on workers’ actual labor market experience Women’s increase in labor market participation my have slowed in 1990’s relative to 1980’s Discrimination narrowed at a slower pace ◦ Concerns about glass ceilings ◦ Pay equivalence in certain occupations may have already been achieved Labor force selection may causes changes in observed wage differentials between men and women. ◦ Wage data only available for labor force participants Identify the causes of the slowdown in the gender pay gap from 1980’s to 1990’s Use accounting techniques to identify… ◦ Changes in the measured characteristics ◦ Changes in the price of measured characteristics ◦ Changes in the unexplained gender pay gap Measure ◦ Effects of unions ◦ Effects of selectivity Identify the size and drivers of the “unexplained” gap Data ◦ Michigan Panel Study of Income Dynamics ◦ Salaried, Non-Agricultural workers (18-65) ◦ Values obtained from 1979,1989, & 1998 data set Unequal lengths in data (10 and 9 years) Compared to Current Population Survey (CPS) Trends remained same in both studies Pay gap fell substantially faster in 1980s than 1990s Female/Male pay ratio ◦ 63% in 1979 ◦ 74% in 1989 ◦ 80% in 1998 Females out-earned males: ◦ 24% in 1979 ◦ 35% in 1989 ◦ 39% in 1998 Pay ratio fell over all earnings levels Price changes over time could affect women either positively or negatively Human capital of women was lower than men and therefore lower returns to experience Technological advances benefited women, in terms of gap, but also increased male wage Union effects and removal Looked at non-agricultural employees ages 18-65 Excluded outliers ◦ Excluded women who made less than $1 and over $250 per hour Looked for the “unexplained gap” ◦ Gap after accounting for differences Gender wage ratio ◦ 70.8% 81.8% 81.2% Relative experience levels resulted in 5.3% decrease in wage gap between 1970-1990 ◦ 1/3 of total decline Gender gap in years of experience decreased from 2.3 years to 1.6 years Deunionization from 80’s to 90’s accounted for 2.4% to 8.9%, which was a 27% decrease ◦ Men moved out of unionized labor Saw that the slowing gap convergence was statistically significant Few women in market place. Percentage of working women fluctuates a lot more than men. Does this cause the gender pay gap slow down to be misleading? Full-Time workers only: ◦ Women: 41-53% Add Part-Time workers: ◦ Women: 68-77% Men: 92-94% Add Workers with wages in PSID ◦ Women: adds 10-13% Men: 80-83% Men: adds 4-5% Make assumptions based on experience and education (above or below the median) ◦ Women: 86-91% Men: 96-98% High percentage of women in labor force were high skilled in the 80’s ◦ Returns to skill ◦ Made gap appear to close faster More low skilled women entered in 90’s ◦ Divorce rates ◦ Single families Selection bias explains 10% of the gap closing from 1979-98 (ceteris paribus) Unmeasured Characteristics 2. Labor Market Discrimination 1. Unmeasured Characteristics ◦ Women’s commitment to labor market increased more in 1980s ◦ Gender gap in housework declined more in 1980s than 1990s 2. Discrimination ◦ 1991 civil rights legislation seems contradictory ◦ Employer’s perceptions changed more in 1980s ◦ Glass ceiling effect 1. Changes in supply and demand were not included in the data Indexes may not be accurate Demand shifted within industries Increase in computerization in 1980s more than 1990s Female / Male Hourly Pay Ratio Procedure - Decomposed Method Results Key findings: ◦ ◦ ◦ ◦ Human Capital Labor Force Selectivity Labor Market Discrimination Supply and Demand Shifts
© Copyright 2026 Paperzz