 |
 |
|||
 |
|
     |
||
| |
|
|||
|
|
|
||
Statistically Speaking
REFINEMENTS IN MEASURING POVERTY 
by Dr. Romulo A. Virola 1
Secretary General, NSCB
It’s the LOVE month once again, and the romantics among us must still be preoccupied with our gifts and how to best celebrate the occasion with our loved ones. Prices of flowers, chocolates and other Valentine gifts have of course escalated to make February 14 all the more memorable. And despite life’s difficulties, we Filipinos still find time, ways and means to express our undying love. Hayyy!!!
But surely, the grandness of the gift or the celebration cannot be a convincing measure of the love one has for his/her partner. Can love be measured?
Like love, there is no standard way to measure poverty. Some people will say that they are poor because they do not have the money to buy a Starbucks coffee or watch the concert of Taylor Swift or watch the Azkals live, particularly the Younghusband brothers, even as Angel Locsin has tweeted Phil to set their date another day! But if you talk to a fisherman residing along the coast of Bataan, their definition of the poor will certainly be different. So, how does one measure poverty?
The United Nations Statistical Commission, which prescribes international standards and guidelines on statistical measurement and practices, has not prescribed any international standard to measure poverty 2. Hence, countries even within the ASEAN region, are using different ways to measure poverty. In addition, while countries use different methodologies to assess poverty, periodic methodological reviews are standard practice among official statisticians in national statistical systems all over the world. This is to ensure that the methodologies and statistical frameworks are updated and reflect current phenomena, taking into account the various developments in the economy and society as well changes in data collection and management systems.
In the case of the Philippines, there have been three refinements in the official methodology for measuring poverty since its adoption in 1987. The first was undertaken in 1992, one major feature of which was the exclusion of alcohol, tobacco, recreation, durable furniture and equipment as well as other miscellaneous expenditures in the list of basic non-food requirements considered in the determination of the poverty line/threshold. Poverty line/threshold refers to the minimum income/expenditure required to meet the basic food and non-food requirements. 3 The NSCB received some flak for this refinement, but for purposes of poverty monitoring, we would not want to include the cost of Red Horse or Marlboro lights in the list of basic needs, would we? Another improvement feature was the use of the consumption pattern of families with income very close to the food threshold instead of families with six members in the estimation of the poverty line/threshold. The food threshold refers to the minimum income/expenditure required for a family/individual to meet the basic food needs, which satisfies the nutritional requirements for economically necessary and socially desirable physical activities. 3 The poverty threshold is indirectly estimated as the food threshold divided by the FE/TBE ratio. 4
In 2003, another improvement was introduced in recognition of the need for poverty estimates with lower levels of disaggregation, specifically at the provincial level. Using the same regional food bundles in the 1987 and 1992 methodologies developed by the Food and Nutrition Research Institute (FNRI), provincial food thresholds were generated using actual provincial prices collected by the Bureau of Agricultural Statistics (BAS) and the National Statistics Office (NSO). With the adoption of this methodology, NSCB was able to release official poverty statistics at the provincial level, which have been very useful in poverty alleviation programs of the government, particularly in the Kapit Bisig Laban Sa Kahirapan - Comprehensive Integrated Delivery of Social Services (KALAHI-CIDSS) program.
The latest refinement in the methodology was approved by the NSCB Executive Board two weeks ago, on 1 February 2011. The review of the methodology was done to address some issues/concerns raised on the estimation of the food threshold, poverty threshold, incidence and other measures of poverty, the relevance of official poverty statistics, and the comparability of poverty estimates across space and over time. It also arose from the instruction of the NSCB Executive Board, in its meeting on 6 November 2002, for the NSCB Technical Staff to undertake a phased revision of the official provincial poverty estimation methodology. The first phase was the one approved per NSCB Resolution No. 1, Series of 2003 Approving the Proposed Methodology for Computation of Provincial Poverty Statistics.
So, how were these latest refinements formulated?
These refinements were based on the recommendations of the Technical Committee on Poverty Statistics (TC PovStat), the group responsible for the development/formulation of the official poverty estimation methodology per Executive Order No. 352 5 known as the Designation of Statistical Activities that will Generate Critical Data for Decision Making of the Government and the Private Sector. This Committee is composed of professionals/experts from the academe, private and government organizations, who have undertaken significant studies on poverty. It is composed of 21 members and chaired by 
Dr. Celia M. Reyes, Senior Research Fellow of the Philippine Institute for Development Studies (PIDS). The NSCB Social Sectors B Division serves as the Secretariat. The complete list of members is shown in Annex 1.
In coming up with the recommendations, the TC PovStat first produced a comprehensive documentation of the various components of the official poverty statistics from its data sources to the generation of poverty estimates, through a project supported by the United Nations Development Programme (UNDP). This Project also reviewed/documented findings from various papers looking into the different issues being raised on the official poverty estimation methodology. The outputs of this Project were then used by the TC PovStat in its overall review, which started in 2008. To facilitate the review, the TC PovStat created four small working groups (SWGs) that focused on the different components of the methodology. The composition of the SWGs is shown in Annex 2.
After more than two years of review and a series of updates to the NSCB Executive Board6, the TC PovStat finalized its recommendations based on the findings of the four SWGs and presented them in a Consultative Forum held on 28 January 20117. The recommendations were then presented and approved during the First Quarter Meeting of the NSCB Executive Board this year. It must also be emphasized that the Committee had earlier decided that if refinements were to be made, it was best to introduce these refinements at the start of a new administration (or at the end of the previous administration) to avoid any misconception that the refinements on the official poverty methodology were politically motivated. Despite the utmost 
professionalism exerted by the TC PovStat and the NSCB Technical Staff and the transparency in the deliberation by the NSCB Executive Board in coming up with the refinements, as almost always, a few unfounded criticisms came up. Ganito na lang ba palagi tayong mga Pinoy?
As mentioned earlier, one major issue that was addressed in the latest refinements is on the comparability of the poverty estimates across space and over time. It has been argued by many poverty analysts that it is important that the yardsticks be consistent across provinces as the poverty statistics are being used by the government in targeting beneficiaries.
Hence, in the estimation of the provincial food threshold, instead of using regional food bundles as what was earlier used, the TC PovStat proposed the adoption of provincial food bundles. It may be noted that as early as 2002, the NSCB Technical Staff has proposed the adoption of provincial food bundles in estimating the provincial food thresholds to capture province-specific characteristics in the estimation of provincial poverty statistics. The provincial food bundles that were adopted have the following characteristics:
“Visualizable” - The food items can actually be cooked together and eaten as a viand;
Nutritionally adequate - Satisfies the 100% Recommended Energy and Nutrient Intake (RENI) for energy and protein and 80% RENI for vitamins and minerals;
Locally available - This is ensured as only the food items that have available prices from the National Statistics Office (NSO) and Bureau of Agricultural Statistics (BAS) were adopted in the food bundle;
Designed based on an indicative national food bundle. The provincial food bundles were developed by starting with a national reference food bundle to enhance consistency across space, and allowing variations in the variety of banana, type of fish, vegetable or snack to consider specific characteristics of the province To illustrate:
Aside from the refinements in the estimation of the food bundle, there were also changes adopted in the estimation of the non-food component. While the food component is measured by using a food bundle, the non-food component is measured indirectly due to the difficulty of determining the required bundle for the non-food items. However, unlike for food where we can tap the experts of the Food and Nutrition Research Institute (FNRI) to define the standard nutrient requirements, for non-food, coming up with standard requirements on education, health, housing, fuel, transportation cost, etc. is a big challenge, not only to the Philippines but also to most countries worldwide. Hence, it is common practice among countries to use indirect estimation for the non-food component. But, while indirect estimation will still be used through the FE/TBE ratio, the approved methodology now uses the average FE/TBE ratio nationally determined from the 2000, 2003, 2006 and 2009 Family Income and Expenditure Survey (FIES) instead of a different FE/TBE ratio for each FIES year. Similar to the adoption of provincial food bundles, it may be noted that the NSCB Technical Staff has also proposed the use of constant/fixed FE/TBE ratio as early as 2002.
But how do all these refinements address the issue of comparability across space and over time?
Well, the FE/TBE ratio is based on the concept of the Engel’s law which is an economic theory that states that the percentage of income allocated to food purchases decreases or goes down as income rises. The FE/TBE ratio is computed from a reference family whose income is near 9 the food threshold to ensure that the non-food requirement is commensurate with the food requirement of a reference family that is appropriate for poverty monitoring. In the old methodology, the FE/TBE ratio of the reference family is computed for each of the provinces with urban-rural disaggregation. In the refined methodology, only one reference family is determined based on the national subsistence incidence in the FIES year. Hence, only one FE/TBE ratio is generated in each FIES year. Further, to ensure that the standard for poverty monitoring will remain the same over time, the average of the 2000, 2003, 2006 and 2009 FE/TBE will be used for at least 12 years or for two administrations. To illustrate, if for example there has been rapid and equitable economic growth, which translated to meaningful increases in the income of Filipinos including the poor, this could result to an increase in non-food expenditure items like pre-paid loads of cellular phones even among the poor. This would therefore result to higher relative expenditure on communication and lower relative expenditure on food, or a lower FE/TBE ratio. Would it then be appropriate to use this smaller FE/TBE ratio for purposes of poverty monitoring? Wouldn’t this be tantamount to raising standards, i.e. losing comparability over time, resulting in a poverty situation that has not changed in spite of the increased income of the poor?
Given all these refinements, how do the estimates based on the refined methodology compare with those using the old methodology?
In general, the resulting poverty estimates based on the refined methodology follow trends consistent with those of the estimates using the old methodology. However, in terms of levels, estimates based on the old methodology were generally higher compared to those generated from the refined methodology.
The 2009 monthly national food threshold for a family of five based on the refined methodology is Php5,298, which is Php429 lower than the food threshold based on the old methodology (Figure 1). Similarly, the national poverty threshold for a family of five in 2009 is Php7,017 using the refined methodology compared to Php7,953 in the old methodology (Figure 2).
Likewise, in terms of subsistence incidence10 and poverty incidence 11 among families, the same trends were observed in both the old and refined methodologies. However, in terms of levels, the subsistence and poverty incidences using the old methodology are higher by 2 and 5 percentage points, respectively, compared to the refined methodology (Figures 3 and 4).
As to be expected, the trends in the subsistence incidence and poverty incidence among population in the two methodologies are also consistent but the levels are higher in the old compared with the refined methodology (Figures 5 and 6).
It must be noted, that poverty incidence among population is the indicator being monitored in the Millennium Development Goals (MDGs), which means that while the target of halving poverty remains the same, the actual level of poverty incidence targeted for 2015 has changed. For MDG monitoring purposes, therefore, the TC PovStat recommended the reestimation of the poverty incidence estimates for 1991 based on the refined methodology. With the reestimation, the Philippine MDG target of reducing poverty is to achieve a poverty incidence among the population of 16.6% by 2015 (one-half of 33.1%) instead of 22.7% (one-half of 45.3%).
And with respect to the magnitude or the count of the food poor and poor families/ population, which is the concern of program implementers like the Department of Social Welfare and Development (DSWD), the trends are also consistent in the two methodologies but the levels under the refined methodology are lower. In particular, the number of poor families estimated as 4.7 million in 2006 in the old methodology, which was the initial target of DSWD in the Conditional Cash Transfer Program is only 3.7 million in the refined methodology. (Table 1)
With the release of the estimates last week based on the refined methodology, a number of “issues”/questions have been raised (Annex 3). We will respond later to these questions/issues in a separate article (under our For The Record page with the following link: http://www.nscb.gov.ph/announce/fortherecord/default.asp ) on our website.
As we celebrate our dates today (hopefully, we all have one), we official statisticians in the NSCB would like to reassure all our stakeholders of our unconditional love for service, integrity and professionalism. Till death do us part!
Happy Valentine’s to all!
Reactions and views are welcome thru email to the author at ra.virola@nscb.gov.ph
_______________
1 Secretary General of the National Statistical Coordination Board (NSCB) and Chairman of the Statistical Research and Training Center (SRTC). He holds a Ph. D. in Statistics from the University of Michigan in Ann Arbor, U.S.A. and has taught mathematics and statistics at the University of the Philippines. He is also a past president of the Philippine Statistical Association. This article was co-written by Bernadette B. Balamban, Statistical Coordination Officer VI of the NSCB. The authors thank Jessamyn O. Encarnacion, Noel S. Nepomuceno, Candido J. Astrologo, Jr., Ma. Libertie V. Masculino, Albert Garcia, Andrea Baylon, and Jeffrey Enrado for the assistance in the preparation of the article. The views expressed in the article are those of the authors and do not necessarily reflect those of the NSCB.
2 While there is no international standard methodology for estimating poverty, the UN Statistical Commission, in its session in 1996 created the Rio Group on Poverty Statistics, which serves as a venue where countries present and compare their experiences and methodologies in measuring and analyzing poverty. In 2006, the Group prepared a Compendium of Best Practices on Poverty Measurement.
3 Per NSCB Resolution No. 2, Series of 2007 – Approving and Adopting the Official Concepts and Definitions for Statistical Purposes for the Poverty Sector
4 The FE/TBE ratio is the ratio of Food Expenditures (FE) to Total Basic Expenditures (TBE) of a reference family, a modification of the Engel’s coefficient.
5 Under EO 352, the Technical Committee on Poverty Statistics (TC PovStat) is designated with the development/formulation of the official poverty estimation methodology while the generation and release of poverty estimates are designated to the NSCB.
6 Updates were presented to the NSCB Executive Board during its meetings in September 2009, November 2009, February 2010, May 2010 and August 2010.
7 A Consultative Forum on the Proposed Refinements in the Official Poverty Estimation Methodology was held on 28 January 2011 at the NSCB Operations Room, Midland Buendia Building, Makati City. The NSCB invited approximately 85 participants from major users and stakeholders from national government agencies, academe, non-government organizations, and other institutions/ organizations.
8 During a Forum on Poverty Lines with Dr. Martin Ravallion held last 3 December 2003, organized by the World Bank – Office in Manila, he said that in general, there was quite a narrow range of issues in the official poverty estimation methodology in the Philippines compared to other countries. He added that there was no one perfect methodology in estimating poverty and that the most important issue that should be addressed by the Philippines was the comparability across space, which could be verified by conducting the Test of Revealed Preference.
9 Families in the +/-10 percentile band around the subsistence incidence.
10 Subsistence Incidence – proportion families/individuals whose income is below the food threshold or whose income is not sufficient to buy food items that will meet the basic food requirements, per NSCB Resolution No. 2, Series of 2007.
11 Poverty Incidence - proportion families/individuals whose income is below the poverty threshold or whose income is not sufficient to buy the basic food and non-food requirements, per NSCB Resolution No. 2, Series of 2007.
Figure 1. Monthly Food Threshold for a Family of Five, Philippines,
2003, 2006 and 2009
Figure 2. Monthly Poverty Threshold for a Family of Five, Philippines,
2003, 2006 and 2009
Figure 3. Subsistence Incidence among Families, Philippines,
2003, 2006 and 2009
Figure 4. Poverty Incidence among Families, Philippines,
2003, 2006 and 2009
Figure 5. Subsistence Incidence among Population, Philippines,
2003, 2006 and 2009
Figure 6. Poverty Incidence among Population, Philippines,
1991, 2003, 2006 and 2009
Table 1. Comparison of Poverty Estimates Based on Old and Refined Methodology
2003, 2006 and 2009
| Province | Old Methodology | Refined | Difference | ||||||
| 2003 | 2006 | 2009 | 2003 | 2006 | 2009 | 2003 | 2006 | 2009 | |
Among families |
|||||||||
Subsistence Incidence (%) |
10.2 |
11.0 |
10.3 |
8.2 |
8.7 |
7.9 |
-2.0 |
-2.3 |
-2.4 |
Magnitude of Food Poor |
1.7 |
1.9 |
1.9 |
1.4 |
1.5 |
1.5 |
-17.6 |
-21.1 |
-21.1 |
Poverty Incidence (%) |
24.4 |
26.9 |
26.3 |
20.0 |
21.1 |
20.9 |
-4.4 |
-5.8 |
-5.4 |
Magnitude of Poor |
4.0 |
4.7 |
4.9 |
3.3 |
3.7 |
3.9 |
-17.5 |
-21.3 |
-20.4 |
Among individual |
|||||||||
Subsistence Incidence (%) |
13.5 |
14.6 |
13.9 |
11.1 |
11.7 |
10.8 |
-2.4 |
-2.9 |
-3.1 |
Magnitude of Food Poor |
10.8 |
12.2 |
12.2 |
8.8 |
9.9 |
9.4 |
-18.5 |
-18.9 |
-23.0 |
Poverty Incidence (%) |
30.0 |
32.9 |
32.6 |
24.9 |
26.4 |
26.5 |
-5.1 |
-6.5 |
-6.1 |
Magnitude of Poor |
23.8 |
27.6 |
28.5 |
19.8 |
22.2 |
23.1 |
-16.8 |
-19.6 |
-18.9 |
Posted 14 February 2011.
 |
 |
|||||||
  E-mail the webmaster |
Terms of Use |
Home • Top of Page | ||||||
1997-2012,
National Statistical Coordination Board
|
||||||||