The value of Self-Rated Poverty
The value of last Wednesday’s report that 50 percent of Filipino families were rated as poor (mahirap) by their heads, in the Social Weather Stations survey of December 2015, is not in that single number, but in its being the latest point in a series of 113 national surveys.
By having the new point, and the Self-Rated Poverty (SRP) percentages of 51 last March, 51 last June, and 50 last September, we know that poverty has been flat in the last four quarters, and that it averaged 50 percent in 2015 (corrected for rounding error).
Given the 2014 SRP quarterly rates of 53 in March, 55 in June, 55 in September, and 52 in December, we already knew, a year ago, that the full-year average was 54 percent in 2014.
The 4-point fall of average SRP from 54 in 2014 to 50 in 2015 is quite significant, since the error margin for a survey of 4,800 households (the pool of four quarterly samples of 1,200 households each) is less than 1.5 percentage points.
The Self-Rated Food Poverty (SRFP) of December 2015 is only 33 percent of households. Given SRFP rates of 36 last March, 37 last June, and 35 last September, this takes the full-year average to 35 percent in 2015. This is a 7-point fall from the average SRFP of 42 percent in 2014. It is a new record low since the SRFP series began in 1988.
The SRP data over time are abundant. The SWS national reporting of SRP was semiannual in 1986-1991. It has been quarterly since 1992, synchronizing the tracking of poverty with that of economic growth.
SRP provides the Philippines with the world’s fastest system of regularly monitoring poverty by means of statistically representative surveys. Its aim is to help understand Philippine poverty over time, not to compare it with other countries.
On the other hand, the official poverty figures over time are scanty. The official data on poverty, based on an official poverty line, had 10 points in the period 1985-2012, due to the old policy of measuring poverty only once in three years. There are official figures for the 2013 and 2014 (first semesters only) also, thanks to a new policy of measuring it annually.
However, there was a midstream change (called a “refinement”) in the official methodology—see “The lowering of the official poverty line,” Opinion, 2/12/11—that has rendered only 7 points comparable. These 7 points are 1991 (to provide a “refined” starting point for the Millennium Development Goals), 2003, 2006, 2009, 2012, 2013 and 2014. The government has not “refined” its figures for 1985, 1988, 1994, 1997 and 2000.
The official figures show that poverty fell over 2012-13, and then rose over 2013-14. The same changes had already been anticipated by the SRP reports—see “Self-rated poverty proves its reliability,” Opinion, 3/14/15.
Based on last year’s experience, the next reference period for official poverty might be the first semester of 2015, when an official survey of income was done, and the report on it might be in March 2016. Will it show that official poverty dropped over 2014-15, and thus confirm the trend in SRP?
Time-series analysis requires consistently-measured data. All of the SWS surveys have used exactly the same SRP methodology as the first-ever national survey, by the Development Academy of the Philippines in April 1983 (when SRP was 55 percent), and the second national survey, by the Bishops-Businessmen’s Conference for Human Development in July 1985 (when SRP hit its all-time peak of 74 percent). The explosion of poverty, obviously due to the hyperinflation in 1984 and 1985, added extreme economic discontent to the political turmoil caused by the assassination of Ninoy Aquino in August 1983.
Econometric analysis of the extensive SRP historical record has found, thus far, that the most important drivers of poverty are price inflation in general, food-price inflation in particular, and underemployment. Growth in GNP as a whole has no discernible impact on poverty, with the possible exception of agricultural sector growth. I think the reason economic growth has not been inclusive is this: It is not matched by improvements in the real wages of most workers.
Self-Rated Poverty is an indicator of subjective wellbeing. Since SRP takes poverty from the viewpoint of the people themselves, it is a bottom-up indicator. Contrast it with poverty defined by an official poverty line, which is a top-down indicator.
Whereas SRP uses the people’s judgements, official poverty uses an institution’s judgement, as to what constitutes poverty. There is no logical inconsistency between 50 percent of the people marking themselves as poor—which is socially meaningful—and 20 percent of them marked as poor by officials using a standard that the people see as stingy. The two numbers have vastly different meanings.
Both SRP and official poverty are measured objectively. The systems of surveying SRP directly and that of surveying income and comparing it to a poverty line are both open to scientific replication. Findings of independent replications will converge.
SRP and official poverty are cross-sectionally similar. Measured by SRP, poverty is always greater in rural than in urban areas. Poverty always falls as the education of the household head rises. Poverty is always greater, the larger the family size.
The superiority of SRP to official poverty is in its amenability to measurement over time. Watching something constantly is basic to learning how it works.
* * *
Contact [email protected].
