Crossan Cooper, Katja Seim & Jintaek Song, A Sin Ban, Not a Sin Tax? Consumption and Incidence of SNAP Soda Restrictions, NBER Working Paper No. 35739 (Sep. 2026):

We study the first statewide SNAP restrictions on sweetened beverages and candy. Restricted spending falls 11 percent; calories and sugar fall 5–8 percent. Substitution is asymmetric: eligible fruit drinks partly offset soda declines, while candy restrictions reduce purchases of eligible snacks. Retailers barely change shelf prices. Losing SNAP’s sales-tax exemption raises tax-inclusive prices only 1.9 percent. Spending declines are larger at retailers with greater SNAP use and remain large online, where checkout stigma is less salient. Nationwide adoption would impose 13 percent of the out-of-pocket burden of an equivalent excise tax, while welfare gains reach roughly $1.1 billion annually.

From the conclusion:

An important question for future research is how restrictions on government income support should be combined with conventional Pigouvian taxes. A restriction tied to an income-support program directly affects only program participants, whereas a Pigouvian tax reaches the broader population. When restrictions leave retail prices largely unchanged, however, they may reduce purchases among low-income households while imposing much less direct financial burden than a tax. This creates a tradeoff between reach and incidence. Future research could examine whether the two policies are substitutes or complements and how the scope of restrictions and the level of taxation should be chosen jointly to reduce harmful consumption while limiting the burden on low-income households.