Abstract:
This paper investigates household consumption changes at retirement by utilizing a comprehensive, diary-based household survey from China. The survey contains both consumption quantity and price information, which permits separating quantity changes from price changes. The mandatory retirement policy in China provides a quasi-experimental setting for identification of the true causal effects of fully anticipated retirement. Using regression discontinuity models, we show that food expenditure declines at retirement, particularly among the low-education group, and that the decline is driven by price declines instead of quantity declines. Shopping time for food increases at retirement, consistent with the price and quantity changes.