Meyer and Sullivan used consumption data, and again they set up the calculation so that the poverty rate for consumption data is the same as the poverty rate for income data as of 1980…By this measure, the poverty rate almost reaches zero percent in 2007, before the Great Recession.
He is referring to a paper by Bruce D. Meyer and James X. Sullivan. Since it comes from a Brookings conference, you can read comments by others at the end of the paper. The commenters did not shoot it down.
The main reason for using a consumption-based measure is that poor people tend to under-report much of their income, such as the value of government-provided benefits. In other words, even if you think that income is the right variable to use for measuring poverty, consumption might be the best available proxy for income.
If the substance of the paper is correct, and poverty is at low levels in the United States, then there is a case for reducing benefits in order to reduce the high marginal tax rate that deters low-income people from working.
However, my own opinion, driven by anecdotal observation rather than data, is that poverty in the U.S. is nowhere near zero. Perhaps if people with low incomes made really good decisions about how to spend their money, then poverty would be near zero. However, over the course of their lifetimes, many people make many bad decisions, and as a result they will spend a lot of time dealing with financial adversity. The moral and practical implications of this view of poverty are not as clearcut as either a progressive or a conservative would like.