The average American pays $36 annually to subsidize food stamps and pays $640 to subsidize corporations.

The claim consists of two parts: the average American paying $36 annually toward food stamps and paying $640 toward corporate subsidies. Multiple sources confirm that the $36 figure for food stamps is approximately accurate for a taxpayer earning around $50,000 per year. For example, Truthout states that in 2012, the average taxpayer making $50,000 paid about $36 toward the food stamps program, which aligns with the claim.

This figure is also referenced in other fact-checking discussions.desmoinesregister.comdesmoinesregister.comHow much does welfare cost the average taxpayer? However, the corporate subsidy figure is more problematic. The claim states $640 annually, but one source from the Des Moines Registerpolitifact.compolitifact.comA Facebook post misleadingly compares how the government spends money collected... mentions corporate welfare costing the average American family about $6,000 a year, which is an order of magnitude higher.

Another sourcedesmoinesregister.comdesmoinesregister.comHow much does welfare cost the average taxpayer? discusses a viral claim of $4,000 in corporate subsidies paid by a $50,000 earner, which is much higher than $640. Politifacttruthout.orgtruthout.orgFood Stamps Are Affordable; Corporate Welfare Is Not notes that such comparisons are often misleading because they conflate different types of government spending and tax expenditures without clear context or consistent methodology. The wide variation in estimates for corporate subsidies arises from differing definitions of what counts as a subsidy and how costs are allocated per taxpayer.

Therefore, while the food stamp subsidy figure is reasonably accurate and supported by data, the corporate subsidy figure of $640 is not well-supported and appears to underestimate typical estimates significantly. The claim as a whole mixes a verified small number with an unclear or underestimated larger number, making it misleading overall.