One thousand statisticians
October is officially National Statistics Month. It was launched at last week’s National Convention on Statistics, with a thousand participants at Edsa Shangri-La Hotel paying registration fees of P1,000 (undergraduates, without meals) up to P6,700 (nonstudents, full board). The convention had 135 technical papers presented over two days, 10 simultaneous sessions at a time. The papers are available on CD; those interested may write to [email protected].
It was refreshing to be among people who understand, from college-level statistics, why a scientific survey of 1,000 respondents is enough to represent the views of many millions of Filipinos. They also know, as professionals or as students aspiring for the field, that Philippine statistics, whether from government or the private sector, are produced by honest practitioners. Statistics is a science, not a branch of advertising.
Predicting a future quarterly from a current monthly or weekly. I was lucky to find a seat in the crowded room where Dr. Bobby Mariano (statistics professor emeritus, University of Pennsylvania) talked about making models for forecasting a future quarterly figure, like GNP, on the basis of current monthly figures of some determinant, like imports.
This is called mixed-frequency modelling. It would be very relevant right now to do a model to forecast the third quarter’s figure of self-rated poverty, considering the rapid increases in the retail price of rice in July and August, as reported weekly by the Bureau of Agricultural Statistics.
Job misery and presidential popularity. In another session, UP statistics professor Dennis Mapa related his main finding that a president’s quarterly SWS satisfaction rating reacts negatively to the same quarter’s “job misery,” which he defined as the sum of the official rates of unemployment and underemployment.
Secondly, he found that the quarterly rating reacts positively to the growth rate of Gross Domestic Product two quarters earlier. This was based on data from the first quarter of 1995 (Ramos period) to the second quarter of 2012 (P-Noy period).
Finding economic status without knowing income. The simple A-B-C-D-E system for classifying Filipino socioeconomic groups has been thoroughly used by business and academe—but not yet by government—for many decades. Everyone knows of some commodities, housing areas, and television programs called D or for the “masa,” and of others called AB or for the upper classes. But the socioeconomic classification system (SEC) is not standardized across research groups; every group—including SWS—has its own house rules, while the government has no rules at all (yet).
A few years ago, the Marketing and Opinion Research Society of the Philippines (MORES), the National Statistics Office (NSO) and former UP statistics dean Lisa Bersales forged an agreement to design a new system for common use by all in the survey industry, including the government. At the convention, Dr. Bersales introduced a new system, 1SEC, that uses a simple two-page questionnaire in order to divide households across nine classes, without asking about income, which currently takes at least four hours of interview time in government surveys. The paper’s discussants (myself included) praised 1SEC, only asking that it be tweaked down to five classes, for practical use.
The value of statistical competition. In my own presentation at the convention, I stressed that, for economists like myself, competition is what guarantees good quality for the public. It is not a zero-sum game. Statistical competition across all sectors, whether government or private or NGO, is a good thing.
Experience from the martial law period taught us that the government won’t get involved, let alone lead, in developing and publishing statistics that might threaten its power structure. Thus, statistics about the state of democracy and governance should be produced by nongovernmental entities.
The government has some statistics on economic deprivation, but these are very meager, due to prejudice in favor of economic growth and against equity of distribution. There are gaps on this subject matter that SWS is trying to fill.
Official poverty statistics will soon be annual. In a corridor, I was able to ask Socioeconomic Planning Secretary Arsy Balisacan (who sits over the government’s entire statistical establishment) about his promise to upgrade the frequency of official poverty data to annual from triennial. He confirmed that the official poverty figure for the entirety of calendar year 2012 will be ready for release later this year. (The official poverty for the first semester of 2012 was released last April 23.)
By now, the government has completed field work for a survey of family income and expenditure in the first semester of 2013, to enable it to report the pertinent poverty figure by 2014. The survey will be repeated next year, to enable release by 2015 of official poverty in the first semester of 2014. Then, in 2015, the survey will be done in both semesters. This conversion from triennial to annual is a great improvement in official tracking of poverty, and is indeed worthy of the convention’s theme: “Statistics that matter to every Filipino.”
* * *
Notice to my brods: Our Pan Xenia 90th Anniversary Dinner that was preempted by the last typhoon will be held starting at 6 p.m. on Friday, Oct. 11, at Manila Hotel’s Champagne Room. See you there!
* * *
Contact SWS: or [email protected].
