Working Papers

On Leave or Out of a Job? The Contract Divide in Employment around Childbirth in Japan

Job Market Paper. Paper here. Slides here.

Abstract: In Japan, most women stop working around childbirth, whatever their contract. Being out of work, however, can mean two very different things: a woman on leave still has a job to return to, while a woman who has lost her job must look for a new one. Most studies of Japanese child penalties measure earnings or employment in ways that cannot tell these two situations apart. Using the Keio Household Panel Survey and the Japan Household Panel Survey from 2004 to 2022, I distinguish the two and follow women through the years after the birth. Regular and non-regular workers stop working at similar rates, but regular workers are mostly on leave, whereas non-regular workers are about 42 percentage points more likely to have lost their job. Women who kept a job are more often back at work in later years, and within couples, mothers account for almost all of the fall in employment. A similar drop in work at childbirth can thus hide very different positions in the labour market, and a count of women in work misses much of what follows for their careers.

Cognitive Ability, Attentional Regulation, and Academic Achievement

Draft in revision; available on request.

Abstract: Research on educational production functions has established that cognitive and noncognitive skills both predict student achievement, but how they combine at the margin remains less settled. Using Growing Up in Ireland, with age-9 baseline controls, I examine how cognitive ability and attentional and behavioural regulation at age 13 combine in predicting Junior Certificate results in Maths and English. The two interact negatively in both subjects: the return to cognitive ability in examination points is larger for students with weaker regulation. The pattern holds when the class teacher rates regulation four years earlier and across checks for multiple testing, sample selection and school clustering. With a broad personality measure of conscientiousness, the interaction is close to zero. Conclusions about how cognitive and noncognitive skills combine therefore depend on what the noncognitive measure captures.

When Do Gender Gaps in Maths Become Skill Gaps? Evidence from Ireland

Draft in revision; available on request.

Abstract: Research on gender differences in Maths achievement has documented persistent gaps, but most studies examine the gap at a single point in time. Using Growing Up in Ireland, I decompose the gender gap in Leaving Certificate Maths with skills measured on the same children at ages 9 and 13, pairing tests of the same domain across ages. With age-9 skills, most of the gap is unexplained. With age-13 skills, most of it is explained, above all by numerical ability. In English, girls' advantage remains unexplained by measured skills at both ages. The same examination gap therefore looks different depending on when skills are measured, and in Maths the skills of early adolescence account for most of it.

Works in Progress