Government definitions of "poor" don't necessarily equate to what you and I would call "poor," due to politics (basically, getting more people identified "poor" helps you get votes if you're the party perceived as better able to help the poor, and in the US one party has dominated the poverty issue and used this strategy for a number of decades).

For example, someone I knew in high school (a number of years ago now) qualified for the school lunch program, so nominally his family was poor. But they lived in a house, with a yard; had cable TV, computers, and Internet. I would say they're working-class, or maybe even lower middle-class.

If the lowest-income 20% of the population is labeled as poor for the purposes of this statistic ("nominal poor"), but the number of people who are so poor their children are starving in the streets ("desperately poor") is 0.20% of the population, then it's not so surprising that statistics over the nominal poor much more closely resemble the general population than the picture of the desperately poor.

During my teen years, my dad made a pretty good software engineer salary, and we lived in a paid off house with a big yard and lots of computers. But with 8 children (and the associated tax exemptions) and substantial charitable donations, we often had net taxable income that was well below various "poverty" thresholds.

Income is a poor approximation of wealth, indeed.

It's the same citation I added to my original post. You'd know this if you bothered to read the post you just responded to.

http://mchb.hrsa.gov/chusa11/hstat/hsa/pages/221oo.html

Read it, I'm not seeing it either.

I'm guessing he's referring to the fraction of underweight children, 5.1% below the poverty line and 5.7% above 4x the poverty line.

Not sure whether 4x should qualify as "rich", but more importantly, being underweight is pretty low bar. For example, I am underweight.

The actual number as reported by USDA is 1%. The keyword to look for is "very low food security", formerly known as "food insecurity with hunger", which is the lowest range of food security measured by USDA. (The second lowest range, "low food security", formerly "food security without hunger", means poor diet but no reduction in food intake. Note that even among the "very low food security" households less than half reported "lost weight" and less than a third reported "did not eat the whole day".)

One interesting observation: "Typically, households classified as having very low food security experienced the condition in 7 months of the year, for a few days in each of those months." I'll go out on a limb and hypothesize that these few days are the last few days of each month. So one thing to try to reduce that number would be making welfare payments daily, instead of monthly. (I read that welfare these is delivered electronically, you carry a special card which looks like a credit card and a certain amount appears on it the first day of each month. So it shouldn't add any overhead to switch to daily installments.)

Sources:

http://www.ers.usda.gov/topics/food-nutrition-assistance/foo... for the definition of food insecurity levels.

http://www.ers.usda.gov/publications/err-economic-research-r... for the 1% number.