1 Kleven et al. (2019)

: Children and Gender Inequality

Reference. Kleven, H., Landais, C., and Søgaard, J. E. (2019). Children and gender inequality: Evidence from Denmark. American Economic Review, 109(2):181-209.  

Research question. How much of the gender earnings gap in Denmark is caused by having children? And how much would remain if no one had children?   

Identification strategy. The paper uses an event-study difference-in-differences design centred on the birth of the first child. For each individual i, define relative time k = t - tᵢ*, where tᵢ* is the year of the first birth. The estimating equation is:   

run separately for men (s=m) and women (s=f), with individual fixed effects αᵢˢ and year-by-education-cell fixed effects γₛₜ. The child penalty at horizon k for women is pₖᶠ = δₖᶠ - δₖᵐ—the differential impact of childbirth on women’s outcomes relative to men’s. This within-person DiD removes time-invariant confounders; the male comparison removes economy-wide shocks.  

Key results. Women’s earnings fall by approximately 20% relative to their pre-birth trajectory in the year of the first birth and remain 20% below their counterfactual ten years later. Men are unaffected. The child penalty accounts for about 80% of the total gender earnings gap in Denmark. Effects are driven by reductions in hours, labour force participation, and occupational switching toward part-time-friendly jobs.  

Takeaway. The gender earnings gap is primarily a child penalty problem, not a labour market discrimination problem per se. Policies that equalise the burden of childcare between men and women—such as non-transferable parental leave for fathers—directly target this mechanism.  

2 Bertrand et al. (2010): Gender Gap Dynamics inHigh-Paying Careers

Reference. Bertrand, M., Goldin, C., and Katz, L. F. (2010). Dynamics of the gender gap for young professionals in the financial and corporate sectors. American Economic Journal: Applied Economics, 2(3):228-255.  

Research question. Why does the gender earnings gap widen so dramatically in the first decade after graduation, particularly in high-paying sectors?   

Data and identification. The paper uses longitudinal data on University of Chicago Booth MBA graduates from 1990 to 2006, matched to detailed employment records. The identification strategy is a human-capital decomposition: within the same cohort and field, compare men and women who differ in (i) MBA specialisation choices, (ii) career interruptions (parental leave, part-time work), and (iii) hours worked.  

Key results. Immediately post-graduation, men and women earn nearly identically (conditional on MBA field and cohort). Within ten years, the gender gap grows to approximately 55%. The decomposition attributes the gap almost entirely to three factors: (i) women are more likely to take career interruptions (time out of work or part-time work reduces earnings by 25-30% per year of interruption); (ii) women work fewer weekly hours on average; and (iii) high-hour jobs in finance and consulting reward long hours nonlinearly—an extra 10 hours per week raises earnings by substantially more than a proportional amount. Unobservable factors and discrimination explain very little of the gap.  

Takeaway. The gender earnings gap in high-paying sectors is driven by the non-linear hours premium and gendered career interruptions, not by discrimination. Policy interventions targeting the value of flexible working arrangements—rather than anti-discrimination law—are more likely to close this gap.  

3 Goldin (2014): The Last Chapter of Gender Convergence   

Reference. Goldin, C. (2014). A grand gender convergence: Its last chapter. American Economic Review, 104(4):1091-1119.  

Research question. What determines the remaining gender pay gap within occupations, after controlling for education and experience? And what would achieve full gender earnings convergence?   

Identification strategy. The paper combines occupational wage data from the US Census, the American Community Survey (ACS), and the NLSY to document within-occupation gender pay gaps and their correlation with the nonlinearity of earnings in hours worked. The identifying variation is cross-occupational: occupations differ in how strongly they reward continuous, long-hours work versus flexible, interruptible work.  

Key results. In occupations where output is divisible and workers are close substitutes for each other (e.g., pharmacy, technology, science), the within-occupation gender pay gap has almost entirely disappeared. In occupations where long, continuous hours are rewarded nonlinearly and client relationships are non-transferable (corporate law, investment banking, medical specialties), large within-occupation gaps persist. The key predictor of the gap is not the share of women in the occupation but the nonlinearity of earnings in hours.  

Takeaway. Full gender earnings convergence requires reducing the premium on long, non-flexible hours—a structural change in how work is organised rather than regulation of employer behaviour.  

4 Olivetti and Petrongolo (2016): The Evolution of Gender Gaps Across Countries   

Reference. Olivetti, C. and Petrongolo, B. (2016). The evolution of gender gaps in industrialised countries. Annual Review of Economics, 8:405-434.  

Research question. How have gender gaps in employment, wages, and hours evolved across OECD countries since 1980? What policies and institutions explain cross-country differences?   

Identification strategy. The paper combines harmonised labour force surveys across 14 OECD countries from 1980 to 2010 with institutional indicators. Cross-country, cross-decade variation in childcare provision, parental leave generosity, and availability of part-time work is used to identify which institutions affect gender gaps.  

Key results. The employment gap between men and women fell from approximately 30 percentage points in 1980 to 15 points in 2010 across OECD countries, but with large cross-country variation. The wage gap fell by 10-15 percentage points. Cross-country differences in employment gaps are strongly correlated with the availability of affordable childcare (reduces the employment gap) and the availability of part-time work (can reduce employment gap but widens hours-adjusted wage gap). Countries with generous non-transferable paternity leave (e.g., Nordic countries) show faster convergence, consistent with the child penalty mechanism of Kleven et al. (2019) .  

Takeaway. Gender gaps are not immutable; institutions and policies that affect the cost of the child penalty can substantially close them. The cross-country evidence provides external validity for single-country event-study estimates.  

References   

1. Bertrand, M., Goldin, C., and Katz, L. F. (2010). Dynamics of the gender gap for young professionals in the financial and corporate sectors. American Economic Journal: Applied Economics, 2(3):228-255.  
2. Goldin, C. (2014). A grand gender convergence: Its last chapter. American Economic Review, 104(4):1091-1119.  
3. Kleven, H., Landais, C., and Søgaard, J. E. (2019). Children and gender inequality: Evidence from Denmark. American Economic Review, 109(2):181-209.  
4. Olivetti, C. and Petrongolo, B. (2016). The evolution of gender gaps in industrialised countries. Annual Review of Economics, 8:405-434.