“House money drives re-betting from winnings: evidence from Japanese bank transaction records” by GAO, Fei (Graduate student, GSE, Waseda), July 7, 10:40-

※English follows Japanese

実証ミクロ経済学ワークショップのお知らせ

日時：7月7日（火）10:40-12:10

場所：３号館12階ディスカッションルーム

発表者：GAO, Fei （早稲田大学大学院経済学研究科博士課程）

タイトル：House money drives re-betting from winnings: evidence from Japanese bank transaction records

要旨：
We use weekly bank-transaction records from a major Japanese bank and two natural experiments to study how gamblers respond to unexpected gambling wins and to a government cash transfer. Gamblers re-bet 7.5% of unexpected wins within one week, and the re-betting persists for about twelve weeks. They barely direct the government transfer into bets, yet spend both windfalls at similar rates on everyday items. After matching on demographics, gamblers spend the government transfer at the same rate as non-gamblers. Gamblers’ response does not differ between light and heavy gamblers. Gambling outcomes shape behavior. When winnings just recover the prior bet, the next bet jumps (break-even effect). A complete loss triggers a spike in non-bet spending instead (loss-coping effect). Bets after losses do not escalate (no loss-chasing). House money, the winnings from gambling itself, is the channel through which gambling sustains itself over time.

担当：別所俊一郎（[email protected])

We are pleased to announce the Empirical Microeconomics Workshop scheduled on 7 July.

Date: Tuesday 7 July 2026

Time: 10:40-12:10

Venue: Discussion room. 12th floor, Building 3.

Speaker: GAO, Fei (Graduate student, GSE, Waseda)

Title: House money drives re-betting from winnings: evidence from Japanese bank transaction records

Abstract:
We use weekly bank-transaction records from a major Japanese bank and two natural experiments to study how gamblers respond to unexpected gambling wins and to a government cash transfer. Gamblers re-bet 7.5% of unexpected wins within one week, and the re-betting persists for about twelve weeks. They barely direct the government transfer into bets, yet spend both windfalls at similar rates on everyday items. After matching on demographics, gamblers spend the government transfer at the same rate as non-gamblers. Gamblers’ response does not differ between light and heavy gamblers. Gambling outcomes shape behavior. When winnings just recover the prior bet, the next bet jumps (break-even effect). A complete loss triggers a spike in non-bet spending instead (loss-coping effect). Bets after losses do not escalate (no loss-chasing). House money, the winnings from gambling itself, is the channel through which gambling sustains itself over time.

Contact: Shunichiro Bessho ([email protected])