I first heard about Charles Karelis’ The Persistence of Poverty when it was published in 2007.  I didn’t just fail to read it; after hearing summaries of its thesis, I considered it too absurd to read.  Now that I’m writing a book on poverty, however, I felt duty-bound to go through the whole book.  When I did, I wasn’t just pleasantly surprised.  I was astounded.  The Persistence of Poverty is an awesome book.  So logical.  So concise.  So direct.  So insightful.  So beautifully written.  While I still don’t buy his big picture of poverty, I have been enriched by the experience.

In honor of this wonderful intellectual product, I’m publishing a seven-part series.  This first post will walk readers through Karelis’ thesis; followups will add critiques and sidebars.  If you have two hours to spare for a mental roller coaster, however, I urge you to actually read the book.

Karelis begins with unwelcome but well-documented facts: A major cause of persistent poverty is the persistently impoverishing behavior of the poor:

Why do poor people often stay poor?  Among the most important causes are five behaviors or, better, nonbehaviors: not working, not finishing school, not saving for a rainy day, not moderating alcohol consumption, and not living within the law.  Obviously not all poor people fail to do these things.  But poor people fail to do them disproportionately.  They account for more than their share of nonworkers, non-school finishers, and so on.  And this contributes to their poverty.

Later in the book, Karelis adds:

One word on a related subject may be in order, before we turn to the third of our five behaviors, failing to save for a rainy day. Having children early and out of wedlock is not among the poverty-causing behaviors we have chosen to focus on. That is because it is not as global or perennial as the other factors on our list. But it is doubtless a big factor in poverty in the United States today.

He could have easily transformed this into a “global and perennial factor” if he phrased it as, “not refraining from having children you are not ready to support.”  In First World countries, this usually takes the form of single motherhood; in the less-developed world and in earlier times, this instead simply took the form of having too many children too early in life.

Karelis then briefly reviews evidence for each of these claims.  For work, education, crime, and savings, it’s clear-cut.  For alcohol (as well as drug abuse), Karelis acknowledges that it’s unclear if the poor drink more in total, but insists that there is strong evidence that the poor are more likely to “binge” and otherwise use intoxicants to the point where they endanger career and family.  For crime, Karelis argues that even “lucrative” crime like theft and robbery usually pays less than minimum wage.  (Never mind all the non-lucrative crimes like brawling, partner abuse, and drunk driving).

What makes these behaviors so puzzling?  Because being poor seems like a strong reason to work hard, excel in school, save your pennies, abstain from alcohol, and obey the law:

Something like the following has to be assumed to reach the conclusion that the conduct in question is inefficient. “Poor people consume little in the way of goods and services. But it is an old and true observation that the less of a good one consumes, the greater the satisfaction one gets from a little bit of it, and conversely that the more of a good one consumes, the less the satisfaction one gets from a little bit of it. Who would deny that a single dollar means more to someone who is struggling to put food on the table than it would mean to her if she had a million other dollars besides? Surely then, the dollars that a poor person can earn by working will be sweet dollars indeed, when compared to the dollars of the rich. So poor people should regard working for pay as very worthwhile. When poor people do not work, they pass up the chance to increase satisfaction on balance. As Lawrence Mead says of non-work, “The puzzle is that poor adults seem less responsive to economic incentives than the better off, even though they need money more.”

“By the same token,” continues the inefficiency argument, “schoolwork, which enhances earning power, is especially valuable to poor people, seeing that they have so much satisfaction to gain by an increase in income. So it is counterproductive for them to drop out of school early, as a disproportionate number do. And heavy drinking by poor people is yet another waste of potential satisfaction, given the extra income that those whose drinking is moderate can expect to earn, and the great satisfaction that is derived from extra income by those whose incomes are small. In a nutshell, the lost income from not working, not building one’s earning capacity though education, and heavy drinking is bound to cost the poor an inordinate amount of satisfaction, and the fact that poor people so often do these things represents a departure from the usual human pursuit of self-interest.

The apparent wastefulness of non-saving and crime rests on much the same logic…

“In fact,” concludes the inefficiency argument, “the wastefulness of non-saving and crime is even greater than this. Failure to save causes not only uneven consumption but reduced total consumption, due to the loss of the investment returns that are normally available from savings. And the cash flow of a career criminal is not only variable but smaller in total than that of a comparable legitimate worker… So there are really two reasons for poor people to save and obey the law. This makes it doubly counterproductive not to.”

Isn’t there a logical explanation?  According to Karelis, there are six standard stories, but none of them explain more than a tiny fraction of what’s going on.  Which is why he need his original resolution of the puzzle.

Coming up: Karelis on why the standard theories fall short, and Karelis’ alternative.


Read Caplan’s next post here.