MEXICO CITY (AP) — Anti-poverty groups in Mexico accused the national statistics agency Monday of arbitrarily changing the way it measures income surveys so poverty appears to be less of a problem.
The statistics agency, known as the INEGI, defended the changes, saying it "improved" the way it measures income because it suspected people were underreporting what they earn.
The agency said it required its interviewers to dig deeper with people who reported no income, to turn up even the most meager sources of income like handouts, odd jobs or help from relatives. Those previously unregistered amounts were then added to the count in the survey, which was carried out late last year.
According to the government poverty agency CONEVAL, the changes implemented last year increased estimates of all household income 11.9 percent nationwide and raised estimates of income among the poorest households as much as 33.6 percent.
But the change makes this year's figures impossible to compare to previous years, frustrating attempts to track or study Mexico's poverty.
The civic group Citizen Action Against Poverty said that changing statistics raised questions about the possible political use of the numbers, and won't change the underlying problem.
"This 'improvement' doesn't change reality," the group wrote in a statement. "People's incomes don't improve because they change the statistical formulas."
According to last year's numbers, 46.2 percent of Mexicans lived in poverty, and 9.5 percent were in "extreme poverty," meaning their income wasn't sufficient to meet their most basic needs.
INEGI said underreporting of income is a problem in many countries, but has been more pronounced in Mexico than anywhere else in Latin America. The agency said it detected the underreporting by comparing reported income to economic activity figures, where it found a difference.
Rolando Ocampo, vice president of the agency's board of directors, said field interviewers were trained to ask additional questions.
"If someone reported no income, we asked them how they got by. They might say, 'My son goes out to work for the woman on the corner,'" Ocampo said, describing an example of previously unreported income.
Some questioned whether such tenuous and unreliable income should be taken into account at all, but Ocampo said it made for a better, more accurate report.
Ocampo said the changes weren't in any way meant to imply poverty rates had improved. But he said that "there was a communication problem" because other agencies and civic groups weren't made fully aware of the changes.