Does banning carbonated beverages in schools decrease student consumption?

I evaluate the effectiveness of carbonated beverage bans in schools by investigating their impact on household soda consumption. I match households in Nielsen Homescan data to their school district's carbonated beverage policies over an eight-year period (2002–2009). I find that when high schools ban the sale of carbonated beverages to students, households with a high school student experiencing the ban increase their consumption of non-diet soda by roughly the equivalent of 3.4 cans per month. I present evidence that this is a substantial offsetting (67–75%) of the average non-diet carbonated beverage consumption in high schools, when these are available to students, thus demonstrating the persistence of preferences when attempting to alter unhealthy habits.