Using Philippine Statistics and Data Sources — A Practical Guide

As of 2026-09-08 · Updated 2026-09-08 · Section: Population, Labor & Education · 日本語版

0. The seven things to grasp first

1. The landscape — which agency publishes what

2. PSA — the core of Philippine statistics

2.1 Legal basis and the incumbent

注 Beware outdated statements of the form "the head of PSA is X." Following the reappointment in May 2024, Mapa remains the incumbent as of August 2026 (sources: Philippine News Agency, "Marcos reappoints Claire Dennis Mapa as National Statistician," 30 May 2024 (reporting the PCO announcement) / Philstar.com the same day / Manila Bulletin, 1 June 2024 = the oath administered by NEDA Secretary Balisacan. The basis for the term is Section 10 of RA 10625 (the National Statistician serves five years and may be reappointed). The official PSA site blocks bots with a human-verification screen, so direct confirmation was not possible).

2.2 Catalogue of the main surveys

APIS and SOF cannot be retrieved via the API (as of 2026-08-25, confirmed by scanning every OpenSTAT node through the API). For OFW matters you have to go separately to BSP's remittance statistics and PSA's SOF.

2.3 Census figures — set the old and new side by side

注(Corrected in the 2026-08-25 cross-check) The previous text had 2015 = 1 May and 2020 = 1 August — the reference dates of two rounds had been swapped. The explanation that "the reference date was pushed back because of COVID" was also wrong; what was pushed back was the start of fieldwork. Ground ① (statutory text): the text of Proclamation No. 1179 (6 July 2021, Duterte) = "I do hereby declare as official for all purposes the results of the 2020 Census of Population, placing the total population of the Philippines as of 01 May 2020 at 109,035,343 persons" (sources: Supreme Court E-Library elibrary.judiciary.gov.ph/thebookshelf/showdocs/7/93607 / LawPhil lawphil.net/executive/proc/proc2021/proc_1179_2021.html). Proclamation No. 973 (11 July 2025, Marcos Jr.) gives a total population of 112,729,484 with a reference date of 1 July 2024 (LawPhil lawphil.net/executive/proc/proc2025/proc_973_2025.html). The 2015 reference date belongs to an intercensal census outside the ten-year cycle prescribed by Batas Pambansa Blg. 72, and PSA states explicitly that "12:01 a.m. of August 1, 2015" was the census reference time and date (PSA, "Highlights of the Philippine Population 2015 Census of Population"; POPCEN 2015 was made official by Proclamation No. 1269 of 13 May 2016). The column headings of citypopulation.de — Census 1990-05-01 / 2000-05-01 / 2010-05-01 / 2015-08-01 / 2020-05-01 / 2024-07-01 — can also be used to cross-check the series, but that is a secondary aggregator, not a primary source. Ground ② (arithmetic using this guide's own figures): the average annual population growth rate of 0.80% below reproduces only when computed over 1 May 2020 → 1 July 2024 = 4.167 years (112,729,484 ÷ 109,035,343 = 1.03388; 1.03388^(1/4.167) − 1 = 0.80%). With a reference date of 1 August (= 3.917 years) it comes out at 0.85%, which does not match the 0.80% in the text. The operating rule in §14 — "if a rate is given, the base value and period can be identified" — works just as well for identifying a reference date. (Settled in the 2026-08-26 source-hygiene pass) The claim that "the proclamation says 109,033,245" was wrong. This guide once noted that "the figure declared in Proclamation No. 1179 is 109,033,245 (2,098 fewer than PSA's published figure), so the number differs by source," but on consulting the text of the proclamation, the proclamation itself says 109,035,343 (as cited above, Supreme Court E-Library / LawPhil). There is no "source discrepancy" of 2,098 people. 注 The figure 109,033,245 is a misprint originating with the English Wikipedia that went into circulation, and must not be used. The 2020 total population is the single figure 109,035,343 (reference date 1 May 2020).

The average annual population growth rate from 2020 to 2024 is 0.80% (PSA). 注 Old figures such as "the Philippine population grows 1.5% a year" are still widely circulated. That was the level of the 1990s and 2000s and is wrong as of 2026.

注 Population figures differ by source:

→ A gap on the order of four million. Use PSA for domestic and administrative purposes and the World Bank for international comparison, and always state the source.

2.4 How the OpenSTAT PxWeb API actually works

PSA's statistical database "OpenSTAT" runs on PxWeb (Nordic statistical database software), and its REST API requires no authentication and is live. Responses were actually confirmed as of 2026-08-25.

Base URL

https://openstat.psa.gov.ph/PXWeb/api/v1/en/DB/

注 api/v2beta/ returns 404 (confirmed 2026-08-25). Only v1 works.

Walking the hierarchy (GET)

The updated field is effectively usable as release detection. For example, updated for the LFS series was 2026-08-05, the quarterly national accounts 2026-08-07, and the CPI 2026-08-05. Even without being able to reach PSA's release calendar, you can catch new publications by hitting the API daily and watching for changes in updated.

The field codes confirmed as of August 2026 (27 fields)

2.5 注 The traps in OpenSTAT (based on actual testing)

3. BSP — external and financial statistics

3.1 What is there

3.2 How to access it

BSP's PXWeb2007 has "PxWeb" in the name but is a different thing from PSA's OpenSTAT and has no REST API. This is easy to confuse.

3.3 注 There are three kinds of remittance figures

This causes more accidents in practice than anything else.

注(Added in the 2026-08-25 cross-check) "Three concepts" is not enough. Even the same concept in the same year differs by estimating body. Personal remittances for 2025 are US$41.560 billion on World Bank WDI vs. US$39.62 billion from BSP — a gap of about US$1.9 billion (roughly 5%). It is not that one is wrong: the World Bank applies its own model adjustments, and its twice-yearly updates rewrite past values retroactively. Practical rule: always write all three of ① the series name (personal / cash), ② the estimating body (BSP / World Bank), and ③ the retrieval date or lastupdated. Never present just one of them as "Philippine remittances." Cross-checking the ratios to GDP: 35.63 ÷ 0.073 = US$488.1 billion; 39.62 ÷ 0.083 = US$477.4 billion. These are broadly consistent with the nominal GDP of US$487.09 billion (2025) in §10.2, so the two BSP-series ratios are internally self-consistent.

Before writing "Philippine remittances are US$X billion a year," always check which concept it is. Personal and cash differ by billions of dollars in any given year, and their growth rates do not agree either.

3.4 注 Foreign reserves also have two concepts

注 The periods and the concepts both differ, so these two must not be set side by side and described as "a decline."

4. Fiscal — the division of labor among DOF, DBM and BTr

注 Do not cite the DOF statistics page as "the latest fiscal data." It has not been updated for nearly two years. Use BTr for fiscal actuals and DBM for budget plans.

注 Do not confuse "the budget amount (GAA/NEP)" with "the actuals (BTr's Cash Operations Report)." The former is a plan; the latter is definitive or preliminary. News reports frequently fail to distinguish the two.

5. Industry and resources — MGB, DOE, DOT, DTI

5.1 MGB (mining)

• Responsible for the value of mineral production, output by metallic and non-metallic mineral, the list of operating mines, and the list of mining tenements (MPSA/FTAA).
• 注 mgb.gov.ph blocks bots with 403 (confirmed 2026-08-25). It must be opened in a browser.
• 注 PSA's OpenSTAT has no substantive mining table (2G/MIN has only ASPBI). MGB is the sole primary source for mineral production figures.
• 注 Claims such as "the Philippines is number X in the world for nickel reserves" or "for gold reserves" appear constantly in industry material, but it is often unclear which indicator the ranking is on (reserves, resources, or production). Where the indicator cannot be identified in a primary source such as the USGS Mineral Commodity Summaries, write "unconfirmed." Also distinguish the claims of the government (MGB) from those of the industry body (Chamber of Mines of the Philippines).

5.2 DOE (energy)

• Energy Balance (primary energy supply and demand), Power Statistics (installed capacity and generation), monitoring of oil and LPG prices, and lists of stations covered by fuel subsidies.
• Under "Data and Price Watch" there are Liquid Fuels Reports and LPG / Auto-LPG Reports.
• 注 The site structure changes frequently, and the statistics themselves are direct Excel/PDF links. 注 The energy field on OpenSTAT (3G) is thin — only power and water.

5.3 DOT / PSA (tourism)

Visitor arrivals (DOT figures obtainable as of 2026-08-25)

注 Provenance of the above (substituted 2026-08-26): the 5,865,319 for 2025 is the DICT eTravel-based figure given in the DOT's annual wrap-up announcement of January 2026 (Secretary Frasco). Sources = PIA (Philippine Information Agency), "DOT caps 2025 with global gastronomy, tourism infrastructure milestones" (pia.gov.ph) / Manila Bulletin, 20 January 2026, "DOT: Quality tourism, not arrivals alone, defined PH gains in 2025" / BusinessWorld, 21 January 2026. 注 The DOT's own statistics page (tourism.gov.ph) could not be reached, so the monthly breakdown and the January–May 2026 figure (2,741,117) are 注 primary source not reached (as of 2026-08-26). 注 2025 was down year on year, so the premise that "post-COVID tourism is rising steadily" does not hold as of 2026.

注 "Tourist numbers" have divergent definitions. Whether it covers foreign visitors only or also Overseas Filipinos makes a difference of hundreds of thousands. Always read the footnotes to the DOT tables.

注(Added in the 2026-08-25 cross-check) The 2025 arrivals figure also splits by "which tabulation system." The 5,865,319 in the table above is on the DICT eTravel basis (the departure/arrival declaration system); on the Bureau of Immigration (BI) basis it is 5,940,975, and total arrivals including Overseas Filipinos is 6,484,060 (DOT). Three figures coexist for the same "2025 tourists," so always state ① the tabulating body (DOT / BI / DICT eTravel) and ② whether it is foreigners only or total arrivals. Sources (substituted on 2026-08-26 for releases closer to primary): PIA, January 2026, DOT annual wrap-up press release = 5,865,319 on the eTravel basis / 5,940,975 on the BI basis (higher because it includes categories such as cruise passengers that eTravel does not fully capture) / total arrivals of 6,484,060 including 543,085 Overseas Filipinos, with provisional international tourism receipts of ₱694 billion (BI basis). Manila Bulletin, 20 January 2026 and BusinessWorld, 21 January 2026 report the same three series. 注 Some outlets, quoting a separate press conference by Secretary Frasco, give 6,430,133 (of which 543,009 Overseas Filipinos), so even the total-arrivals figure has two versions in circulation. The DOT's own statistical tables could not be reached (注 primary source not reached).

5.4 DTI / BOI / PEZA (investment)

• "Approved investments" are not amounts actually invested. BOI and PEZA approval-basis figures are reported as "investment up X%," but there is no guarantee that the projects break ground or operate.
• For money that actually came in, look at BSP's FDI (BOP basis, net inflow). The two diverge widely every year.
• PSA's OpenSTAT also has a 2B/FI (Approved Investment) node.

6. Education — DepEd, CHED, TESDA

注 All three blocked bot access with 403 as of 2026-08-25 (deped.gov.ph, ched.gov.ph, tesda.gov.ph). Automated collection is impossible; manual browsing is assumed.

注 The education field on OpenSTAT (1C) has only four tables (measured 2026-08-25): - Public and private enrollment (2000–2024) - Net enrollment rate for kindergarten - Number of public school teachers (SY 2004-05 to SY 2022-23) - Household spending on education (FIES basis: 2006, 2009, 2012, 2015, 2018, 2021, 2023)

→ PSA is not sufficient as the primary source for education statistics. For deeper figures go directly to DepEd/CHED/TESDA.

注 School Year (SY) notation spans two years, as in "SY 2025-2026." Do not join it directly to calendar-year economic statistics.

7. Labor and wages — DOLE and NWPC

7.1 Employment statistics come from PSA; the institutional side from DOLE

• The unemployment rate, the number employed and the underemployment rate come solely from PSA's LFS (monthly). DOLE handles policy and institutions.
• OpenSTAT holds a long LFS series from April 2005 to June 2026.
• 注 The LFS definitions changed in October 2021: the industry classification moved from the 1994 PSIC to the 2009 PSIC and the occupational classification from the 1992 PSOC to the 2012 PSOC, so the tables are broken in two. They are separate tables on OpenSTAT as well:
• Employment by industry: 0091B3FEMP1.px (April 2005 – July 2011, 1994 PSIC) / 0101B3FEMP2.px (January 2012 – June 2026, 2009 PSIC)
• Employment by occupation: 0111B3FEMP3.px (April 2005 – January 2016, 1992 PSOC) / 0121B3FEMP4.px (April 2016 – June 2026, 2012 PSOC) → They must not be joined into a single long time series.

7.2 NWPC (minimum wages)

• The Regional Tripartite Wages and Productivity Boards (RTWPBs) issue Wage Orders region by region. They are not uniform nationwide but differ by region, sector (agricultural / non-agricultural) and business size.
• 注 The NWPC site's "Current Real Minimum Wage Rates" page still linked to a PDF dated September 2022 as of August 2026 (verified). Do not trust the word "Current."
• Nominal minimum wages are revised frequently, Wage Order by Wage Order. Always record the Wage Order number together with the effectivity date.
• 注 Distinguish nominal from real wages. Reporting usually deals only with the nominal figure.

8. Disasters — NDRRMC, PAGASA, PHIVOLCS

8.1 Requesting PAGASA's raw data goes through a "Google form"

The practical workflow as confirmed on 2026-08-25:

There is also a status-inquiry page for tracking progress. 注 It is not a download service, so allow for lead time.

8.2 注 PHIVOLCS has a broken SSL certificate chain

As of 2026-08-25, phivolcs.dost.gov.ph and earthquake.phivolcs.dost.gov.ph could not be retrieved automatically because of an "unable to verify the first certificate" error — the intermediate certificate is not being served correctly. 注 Disabling certificate verification when scraping is dangerous, so retrieve manually, or use the USGS earthquake catalogue (which has an API) as an alternative.

注 Typhoons have two names — an international name (Typhoon X) and a domestic Philippine name (the PAGASA name). Domestic reporting uses the latter. Without both, cross-referencing is impossible.

9. Elections — COMELEC

注 comelec.gov.ph blocks bots with 403 (confirmed 2026-08-25). Canvass results are mainly PDFs and images; no machine-readable CSV is officially provided. If you use data reprocessed by civil-society groups, always reconcile it against the COMELEC originals.

注 The terms of the legislators elected in the 2025 midterms run to 30 June 2028. Bear that in mind when writing about "sitting legislators."

10. International agencies — which ones are usable

10.1 Comparison of usability (measured 2026-08-25)

10.2 A worked example of the World Bank API (verified working)

https://api.worldbank.org/v2/country/PHL/indicator/NY.GDP.MKTP.CD?format=json&date=2022:2025

Values actually retrieved on 2026-08-25 (lastupdated: 2026-07-13):

Key indicator codes: NY.GDP.MKTP.CD (nominal GDP), SP.POP.TOTL (population), BX.TRF.PWKR.CD.DT (remittances received), FI.RES.TOTL.CD (total reserves), NY.GDP.PCAP.CD (GDP per capita).

注 Always check lastupdated. The World Bank makes large updates about twice a year, and past values are rewritten retroactively.

10.3 UN Comtrade — the "preview" API you can hit without a key

A worked example (successfully retrieved 2026-08-25):

https://comtradeapi.un.org/public/v1/preview/C/A/HS?reporterCode=608&period=2024&flowCode=X&partnerCode=0&cmdCode=TOTAL

→ Total 2024 exports (FOB) of the Philippines (reporterCode = 608) = US$73,268,718,497

注 This response carries "isReported": false. "TOTAL" is a derived value that Comtrade computed by summing commodities; it is not a total the Philippines reported directly. It differs from PSA's published International Merchandise Trade Statistics (IMTS) in rounding and revision timing.

注 The preview API has record limits and rate limits (the specific numbers are not stated in the official documentation). For serious retrieval, get a free subscription key.

11. 注 A catalogue of traps

11.1 Changes of base year

注 Do not set 2018-base GDP alongside 2021-base real income in the same document. Without stating the base year, readers will misunderstand.

注 The quarterly national accounts cover 2000 to 2026, while the "Annual GDP Series" covers 1946 to 2023 in a separate node, built differently. Do not casually join them when making a long-run chart.

11.2 Old and new region codes (the most important trap)

注 Practical consequences 1. The figures for Western Visayas (Region VI) and Central Visayas (Region VII) are discontinuous across the NIR separation. Do not run time-series comparisons across it as they stand. 2. Whether there are 17 or 18 regions differs by source. Beware older material that says "the Philippines has 17 regions." 3. When joining on codes, always record which PSGC vintage — old or new — you are using. Joining across vintages produces missing values. 4. Barangays are also created, merged and promoted to cities every year. The PSGC is a living thing, updated quarterly.

注 In addition, there have been recent legal changes over the constituent LGUs of BARMM, and the affiliation of provinces can differ by source. For any BARMM-related regional aggregation, always check the reference date of that material.

11.3 The "PDF-only" problem

Always have two people cross-check transcription from PDF tables. Missing a digit grouping, a unit (thousand / million / billion pesos) or a footnote (p = preliminary, r = revised) is fatal.

11.4 Bot blocking (measured 2026-08-25)

For PSA, "the main site is 403 but the API is alive." Knowing this changes working efficiency by an order of magnitude.

11.5 data.gov.ph is dormant

As of 2026-08-25, accessing data.gov.ph returns an essentially empty page with only a title, and hitting the CKAN-standard /api/3/action/package_list returns no dataset list.

Starting work on the assumption that "the Philippine open data portal = data.gov.ph" wastes time. The real entry points for open data are PSA OpenSTAT's PxWeb API and the individual departmental sites.

11.6 The etiquette of revisions

注 For quarterly GDP, the previous quarter is revised when the following quarter is published — this is the norm. It is normal for "the value at the time of announcement" and "the value now" to differ. When citing, write "the value as published on [date]."

12. Criteria for deciding "which figure to use"

12.1 Principles of precedence

12.2 Recommended sources by purpose

12.3 A checklist before writing "number X in the world"

注 Ranking claims are the single most error-prone category in writing about the Philippines. Do not write one unless all of the following are satisfied.

Likewise, standard claims such as "the BPO industry is second in the world," "shipbuilding fourth in the world," "second in the world for nickel reserves" should be checked every time for whether the indicator, the year and the source can be identified.

13. A practical checklist

Run every single figure you write through the following.

Recording the access date is mandatory. Philippine statistical sites change their URL structures frequently, and it is not unusual for links to break within six months.

14. Appendix: endpoints you can use immediately (verified working 2026-08-25)

[Related] Reading the PSGC and the regional divisions / Reading BSP monetary policy and the exchange rate / Philippine public finance and the budget process (GAA, NEP) / The labor market and the minimum wage system / The institutions and figures of the mining sector

Cross-check record (2026-08-25)

Assigned cluster: a cross-cutting audit of all figures (a mechanical grep sweep across all 76 documents)

Cross-cutting indicators covered: population / GDP growth / inflation / the peso rate / the policy rate / minimum wages / electricity prices / remittances / number of regions / number of provinces / number of islands / number of Filipinos resident in Japan.

A. Discrepancies detected and how they were handled

B. Cross-cutting indicators confirmed to agree (no discrepancy)

C. The operating rules that came out of this audit (five)

1. Write the year whenever you write "how many provinces" or "how many regions." The number of provinces went from 81 to 82 in 2022 and the number of regions from 17 to 18 in 2024. Citations that straddle the year an institution changed are the most accident-prone.
2. For penetration rates and shares, always write the provenance of the denominator. The UN estimate (117 million) and the definitive census figure (112,729,484) move the result by several points.
3. Remittances need the three-part set of series name + estimating body + retrieval date. Personal vs. cash differ by about US$4 billion, and BSP vs. the World Bank by about US$1.9 billion.
4. Do not go back and forth between pesos and dollars. For GDP the price basis (real vs. nominal) differs, and for project costs the rate differs with the date of announcement. The moment you convert, it becomes a different number.
5. Work backwards from a growth rate to check the base value. The only reason this audit could pin down the third UCME figure was the single division +65.5% = 0.4405 ÷ 0.2662. If a rate is given, the base value can be identified.

Cross-check record (2026-08-25, second round: a cross-cutting audit of all figures)

Assigned cluster: a cross-cutting audit of all figures (a second grep sweep across all 76 documents)

The first audit (table A, #1–#11 in this section) uncovered institutional values (numbers of provinces and regions) and definitional differences (remittances, denominators, price bases). The theme of the second audit was different. What it found was "preliminary values left in place" and "the obsolescence of the cross-check records themselves."

D. New discrepancies detected in the second round

E. Cross-cutting indicators re-confirmed to agree in the second round (additions)

F. The operating rules that came out of the second audit (three)

6. "NDRRMC" alone does not identify a disaster statistic. Always write the date of the sitrep. The largest source of discrepancy in this audit was that three documents, all citing "the NDRRMC," were scattered across an early-November flash figure, a 12 November interim figure, and a late series. The number affected multiplies several times over in the two weeks after an event.
7. Once a matter is settled, always rewrite the record that said "unconfirmed." #18 was a new species of discrepancy: the cross-check record had become older than the body text. A cross-check record is part of the material and is subject to freshness management.
8. At the end, always reconcile a document's opening summary against the notes in its body. #19 was a case of asserting in one's own opening a figure one had oneself described as uncorroborated. Summaries are easily left behind, unrewritten, after the body has changed.

In sum: the first round dug out "differences of definition," the second "differences of point in time." Institutional values (the number of provinces, the number of islands) are stable once fixed, but figures for disasters, prices and exchange rates are out of date again the week after you fix them. That is why what this body of material really needs is not "correct figures" but a state in which every figure carries a date and a provenance.

日本語の詳細版: 分野「統計・データ」「フィリピン統計とデータソースの使い方 ― 実務ガイド」（本サイト日本語版）。関連＝「PSGC と地域区分の読み方」「BSP 金融政策と為替の見方」「フィリピン財政と予算プロセス（GAA・NEP）」「労働市場と最低賃金制度」「鉱業セクターの制度と数字」。