Working papers

Toward Better Informed Decision-Making: the Impacts of a Mass Media Campaign on Women's Outcomes in Occupied Japan (2020, submitted)

This study examines the impact of women's radio programs that the US-led occupying force aired in Japan (1945-1952) to dismantle the prewar patriarchal norms. Using local variation in radio signal strength driven by soil conditions as an instrumental variable, I provide causal evidence that greater exposure to women's radio programs increases women's electoral turnout, which further translates into a greater vote share for female candidates. This positive effect contributes to women's greater representation in the national legislature: had there not been women's radio programs in place, the number of female winners would have been halved. Moreover, exposure to women's radio programs contributes to a decline in fertility and therefore had an important implication for the nation's demographic landscape. The declining fertility is due neither to an increase in women's career aspirations nor to a decline in marriages. My results are not driven by a preexisting correlation between radio signal strength and women's behavior before the US occupation. Although research shows that gender norms have historical roots and move slowly, my findings provide evidence that public policy can cut against them to promote equal participation in decision-making.

The Consequences of Increased Women's Representation in Politics: Evidence from Japan (2018)

The female share of the Japanese national legislature began trending upwards in the mid-1990s, after 50 years of hovering around three percent. To better understand the consequences of this change, I examine the impact of female office holding on electoral and public policy outcomes in the context of the Japanese House of Representatives in the last two decades (1996-2017). To isolate the causal effect of legislators' gender, I employ a regression discontinuity design in a sample of contests where a female candidate wins or loses against a male candidate by a narrow margin. As opposed to the long-standing argument that female electoral victories lead more women to run for office, an additional female victory does not affect the number of female candidates nor female winners in the next election. This result is consistent with Ferreira & Gyourko's (2014) and Broockman's (2014) findings for US mayoral and state legislative elections. Moreover, to examine whether additional female officeholding affects legislative outcomes, I turn to patterns of petitions signed by each legislator. Unlike roll-call votes, legislators do not necessarily act along party lines when considering petitions. I find that female legislators are more likely to support petitions on subjects supported by women, namely gender issues and reproductive rights. Therefore, as models of identity politics would predict, the presence of female legislators is an important channel through which women express their needs. Furthermore, contrary to a wide-spread concern that women are less qualified as politicians than men, the gender of legislators does not affect the amount of per-capita intra-governmental transfers, which provides one piece of evidence that female legislators are as good as male legislators at serving the interests of their local constituents.

Intrahousehold welfare (2019, with Pierre-André Chiappori and Costas Meghir)

Measuring individual-level well-being is crucial when we think about public policy, and existing literature has shown that substantial inequality is hidden within the household (Lise & Seitz, 2011). Earlier work has addressed such intra-household inequality, but it has yet to account for household public goods when measuring it. To shed light on this issue, we develop a new way to measure intra-household inequality that reflects individual well-being derived from household public goods. Using a collective household model, we defined the Money Metric Welfare Index (MMWI), which is essentially a monetary amount that a person would need to reach the current utility level if she were to pay the full price of household public goods. We show that the MMWI is uniquely identified up to an increasing transformation. Then we structurally estimate the collective household model and compute the MMWI using the Japanese Panel Survey of Consumers (JPSC). The JPSC is particularly suitable for our purpose because it provides variables that the MMWI requires, including expenditure for household public goods and time use. Estimated MMWI allows us to see how the intrahousehold inequality evolved in the last 20 years.


“Land and Real Estate Price Sensitivity to a Disaster: Evidence from the 2011 Thai Floods"

(with Yasuyuki Sawada, Hiroyuki Nakata, and Kunio Sekiguchi (2018), Economic Bulletin, AccessEcon, vol. 38(1), pages 89-97)

“The Long-Run Socio-Economic Consequences of a Large Disaster: The 1995 Earthquake in Kobe”

(with William duPont IV, Ilan Noy, and Yasuyuki Sawada (2015), PLoS ONE 10(10))

Working in Progress

Causes and consequences of the sex segregation in college major (with Mika Akesaka)