At the Manhattan Contrarian blog, one of my favorite themes is the government's "poverty" scam—that is, the obvious fraud by which the government ridiculously inflates the measured "poverty" rate by refusing to count hundreds of billions of dollars of government spending as alleviating poverty, and by willfully ignoring huge sources of income. And then the government uses the fraudulently inflated poverty rate to justify continuing and increasing some $1 trillion in annual anti-poverty spending. For an introduction, two representative articles are Poverty: It's Worse Than You Think and The Poverty Deception, Part II . 

What never ceases to amaze is that literally everybody gets taken in by the scam. By "gets taken in," I mean, treats the government "poverty" numbers as some kind of legitimate effort to get a handle on suffering and deprivation in the country, as opposed to the obvious fraud that they are. And by "literally everybody," I mean, everybody up to and including scholars at think tanks who make a living studying the numbers and who just don't have any excuse for not knowing better. (I would give Robert Rector of the Heritage Foundation as an exception, but a lonely one.)

As the latest example of the genre, we have on yesterday's Wall Street Journal op-ed page an article by Robert Doar of AEI titled "Mismeasuring Poverty." Doar calls attention to some recent scholarship by Bruce Meyer of the University of Chicago and Nikolas Mittag of Charles University that indicates in at least a couple of respects that the Census Bureau may be overstating "poverty" somewhat:

Here’s good news for policy makers—on the right and left—concerned about poverty in the United States. A new study by economists Bruce Meyer of the University of Chicago and Nikolas Mittag of Charles University shows that public-assistance programs are far more effective in alleviating poverty than many government statistics suggest.

Doar reports that the Meyer/Mittag research has uncovered that the Census poverty data, derived from its Current Population Survey (CPS), miss large amounts of government handout benefits that, if counted, would lower reported poverty substantially. In a review of data just from the state of New York, the researchers found that the Census CPS data missed entirely a third of housing assistance recipients, 40% of SNAP (food stamps) recipients, and 60% of TANF (cash welfare) recipients. Oh, and then, even for those recipients that were identified, the CPS responses underreported the value of the benefits by 6% for SNAP, 40% for cash and 74% for housing assistance.

Now these are rather gigantic omissions—just these omissions have to come to at least $100 billion per year on a national basis—and you would think that I might applaud somebody pointing this out. Instead, my reaction is that the whole article and the underlying research are touching in their absurd naivete. A few points:

  • As I stated in Poverty: It's Worse Than You Think, the most fundamental flaw of the whole Census poverty-measuring game is that they just send out survey forms and accept whatever the recipient sends back without any kind of double-check or verification. I wrote then: They take whatever answer a person gives as to his or her household income, without any sort of follow up or double check. For that article, I interviewed Trudi Renwick, the Chief of the Poverty Statistics Branch of the Census Bureau, and yes, that's what she told me. And that's not just as to their government benefits, it's as to their income as well.
  • Of course people at the lower end of the income spectrum lowball their income and benefits received when they answer this survey. They know that the lower your income/benefits are the more additional benefits you may be entitled to; and they don't know whether this form may be used to audit or reduce their benefits.
  • While it may be difficult or (in some cases) even impossible for the government to check on whether people correctly report their income, they can definitely check on whether people correctly report their benefits. And, even if they don't do that individually, they can easily calculate the aggregate level of benefits in each category for the country as a whole implied by their survey results and compare that to the government's known budget expenditures. In fact, if they don't do this obvious double-check, they are either completely incompetent, or (what I am sure is the case) fraudulent. The Meyer/Mittag study implies that the CPS survey is capturing at best half of government benefit payments. Absolutely clearly, Census has known this for years or decades and has been willfully suppressing it.
  • How could you discover a gigantic discrepancy like this in the reporting of government benefits and not go on to ask the next obvious question, which is how much is income being underreported at the low end? As I again pointed out in Poverty: It's Worse Than You Think, there is every reason to think that the CPS survey misses almost all, if not all, of the underground economy. I mean, if you were getting paid off the books as a construction worker or nanny or housekeeper—or for that matter, as a drug dealer—and you got one of these CPS forms in the mail, are you really going to just put down the accurate amount of your illicit income and hope the government doesn't use that against you? The whole idea is ridiculous. So how big is the underground economy? A 2011 University of Wisconsin study to be found at the link puts it at about $2 trillion per year, around 12% of the economy. And there's every reason to think that the lion's share of this money goes to people who show up as low or no income on the CPS survey.
  • And then the Doar article somehow just doesn't discuss that most government benefits are intentionally not counted in the measure of poverty as a result of contrived definitions by which they are deemed not to be income. Of the benefits discussed in the Doar article, only TANF gets counted in the measure of poverty. The list of benefits intentionally omitted runs from SNAP to housing assistance to Medicaid to EITC to energy assistance and on and on. The amount of benefits intentionally omitted from the definition exceeds the figure that would be sufficient to eliminate all poverty in this country by a factor of 2 to 3.

So to the $64,000 question: If you make the corrections to the Census CPS data implied by the Meyer/Mittag study and add the missing benefits to the reported income of the recipients, how much does the poverty rate move? The answer is not in the Doar article. How could he leave it out—or not at least raise the question? And let me further raise this question: Suppose you add the Meyer/Mittag missing benefits to the CPS data, and also add the missing underground economy, and also add the value of the intentionally uncounted benefits like SNAP, EITC, housing assistance and Medicaid. What is the poverty rate then? The most recent fake data from the Census Bureau gives the official "poverty" rate as 14.8%, but with these corrections you'd be very hard pressed to get it to 2%. Literally everybody you think of as being poor would be removed by the counting of the benefits, except for a few living like the Unabomber in a cabin in the woods. The bulk of the people who would be left in "poverty" would be oddball cases you would never have thought of, and would never think of as "poor," like graduate students at Harvard living on fellowships (they don't count either).