Structure, bases, sources and revision policy for the aseanx FDI Pulse. This note governs every edition; figures for a given reference year are published separately.
This note covers the FDI Pulse only. The Business Conditions Index is a separate instrument with its own method note: it is compiled on a different basis and, from the next cycle, on a different calendar. Sections that govern both instruments — purpose, the coverage rule, bases, sourcing and revision policy — are repeated in each note rather than split across a third document.
aseanx publishes two analytical instruments covering the eleven ASEAN member states. The Business Conditions Index summarises recorded business conditions in each market as a single score. The FDI Pulse describes where foreign direct investment has landed, in what composition, and with what stability.
Two principles govern both, and the remainder of this note follows from them.
Traceability. Every published figure originates with a named statistical authority, central bank, investment board or multilateral institution, carries the date on which it was obtained, and can be verified at source. No figure is an estimate, a model output or an interpolation. Both instruments compute from published input files through published calculations; applying the calculation to the inputs reproduces the published result.
Disclosure of gaps. Where an issuing body does not publish a measure, the absence is shown. Gaps are not filled with regional averages, neighbouring-market proxies or estimates. The distribution of what is measurable across the eleven markets is treated as information in its own right.
Both instruments are descriptive. They summarise what was recorded in a defined reference period. Neither forecasts, and neither should be read as a forward-looking assessment.
A series is used in a regional statement only where one issuing body compiles it for all eleven markets on the same definitional basis. Series with narrower coverage are reported for the markets that publish them and are not carried into any regional figure.
The rule exists because the Pulse is a comparative instrument. A regional figure assembled from different definitions in different markets would not support comparison between them, whatever its internal consistency. The rule is applied symmetrically: were a series to cease publication for any one market, it would leave the regional statement for all eleven rather than be estimated for the market concerned.
Where a market does not report a series, the gap is shown as a gap on that market's own row. It is never filled with a regional average, a neighbouring market's value or a carried-forward figure.
The Pulse reports a distribution rather than a level. Each market carries five defined readings and the region carries three.
A single score requires components that are commensurable — measured in the same way, on the same basis, so that adding them means something. The Pulse’s components are not. They comprise net cross-border flows, gross investment-board pledges and accumulated stock positions: three distinct kinds of measurement. Combining them into one figure would assert an arithmetic relationship that the underlying data does not support.
Two further properties would compound the effect. Such a figure would be dominated by the largest recipient market, which accounts for the majority of recorded regional inflows and is itself subject to the double-counting limitation set out below. And it would rank proportional volatility above scale, so that a small movement on a very small base would outrank a substantial movement on a large one.
The absence of a composite is therefore a methodological position rather than an omission. The question the Pulse answers concerns the shape of a distribution, and a distribution is described rather than averaged.
Four classifications are applied: small and lumpy where the ten-year mean is below US$1 billion or two or more years were net-negative; swinging where the coefficient of variation exceeds 0.6 or one year was net-negative; variable where it exceeds 0.35; and steady otherwise. The thresholds are conventions. The coefficient of variation and the ten-year series are published beside every classification, and the ten-year mean and negative-year count are printed for each market whose classification turns on them, so that the small-and-lumpy test can be checked rather than taken on trust.
The classifications are independent of the direction reading by construction. A market may be classified steady and record a substantial fall in its latest year — a movement within a stable band. A composite would merge the two and eliminate the distinction.
Investment figures are compiled on several bases which are not interchangeable, and the Pulse quotes each on its own basis without conversion.
Because the bases differ, national authorities and balance-of-payments compilers publish materially different figures for the same market and year. Each edition carries a discrepancy register recording every such gap with the reason for it. The register is not a reconciliation and does not attempt one; it is the disclosure required to use either figure correctly.
Market-level figures for ASEAN are compiled by the ASEAN Secretariat in the ASEAN FDI Database and published in the ASEAN Investment Report; world and cross-regional comparisons appearing in the same report are compiled by UNCTAD. Within the Database, member states submit through different national bodies — central banks in seven markets, statistical authorities in two, a joint survey in one, and, in Myanmar's case, the investment and company registrar.
Where the Secretariat's compilation differs from the World Bank's balance-of-payments series for the same market and year, three documented causes apply: presentation basis, where members submit on a directional basis while publishing their own accounts on the asset and liability principle; manual vintage, where a member submits on an earlier edition of the balance-of-payments standard by agreement; and reporting vintage, submissions being preliminary in the first instance and final typically two years after the reference period. No line-by-line reconciliation has been published by either body, and none is asserted.
Every input carries one of three verification levels, recorded in the published data and shown on the page where the distinction affects how a figure should be read.
Figures at secondary level are identified as such wherever they appear. Where primary access is subsequently obtained, the verification level is updated and the change is recorded, whether or not the figure itself changes.
A quarterly companion covers within-year movement. Merchandise trade is the only measure published at monthly or quarterly frequency by an issuing body in all eleven markets, and it is the basis of the quarterly reading: the year-on-year change in the eleven markets' combined merchandise trade. Quarterly gross domestic product (GDP) is compiled in seven markets and is reported as a seven-market reading, never aggregated to eleven.
The publication date is set by the slowest reporting authority in each cycle, which places it at the fifteenth day of the third month following the close of each quarter.
The FDI Pulse follows its principal sources. Balance-of-payments figures update as the compiling authority publishes; the regional sector and source composition updates with each edition of the ASEAN Investment Report, which has been published between September and December in recent years.
Where an authority has not yet published, the figure for the preceding period is shown with its true reference period and the date on which the new figure is expected. It is never re-dated. A regional aggregate is published only when all constituent markets have reported; if published earlier it is marked provisional with the number of reporting markets stated, and reissued when complete.
Revisions by a compiling authority flow through the published calculation without adjustment; every input records the date on which it was obtained. A change to a component, a weight, a multiplier or a threshold is a change to this note, is published as such, and is accompanied by a restatement of the affected prior figures. Corrections to a published figure are issued with the date of correction and a statement of what changed.
Circumstances that would alter published figures include: a revision by a compiling authority; the arrival of a current-year value for a market previously a year behind; the discontinuation of a core series for any market, which would remove that component from the core for all eleven; the first compilation of a series for a market currently outside coverage; and the publication of a regional series meeting the coverage rule, which would be assessed against the same test.
The FDI Pulse publishes no composite figure. The absence is a methodological position, not an omission.
A single score is only meaningful where the components behind it are commensurable — each measured in the same units, on the same definitional basis, so that combining them means something. The Pulse’s components are not. The evidence it reads comprises three different kinds of measurement:
A single figure computed across those three would assert an arithmetic the underlying data does not support. The point holds inside Table 1 as well: UNCTAD footnotes Brunei Darussalam and Viet Nam as asset/liability and the Philippines as directional calculated from asset/liability, so even one issuer's own table is not one presentation basis throughout.
A composite would also be dominated by the largest recipient — Singapore alone is 61.8% of recorded inflows — and it would reward volatility over scale: Brunei Darussalam's 546.2% rise, on a base of US$26 million, would outrank every other movement of the eleven.
The question the Pulse answers is a question about a distribution. A distribution is reported by describing its shape, not by collapsing it to a mean. Each market therefore carries the readings its source supports and no more — level, direction and share of the regional total — and the region carries one more, concentration. Where a reading is not published for the reference year, Section 2 records it as not published rather than filling it from an older vintage.
Two consequences follow. A percentage change is never reported instead of the level it is computed on. Brunei Darussalam's inflows rose 546.2% in 2025, from US$26 million to US$168 million, which is one tenth of one per cent of the region: read the level, not the rate. And direction is never allowed to stand for weight. Singapore rose 10.8% and Lao PDR rose 7.8%; the first of those two movements is US$14.7 billion and the second is US$0.1 billion. A composite would merge those readings and eliminate the distinction.
Three checks are applied to the market table before any edition is published. The eleven market rows must sum to the printed regional total, allowing only for rounding. The table must produce the number of declining markets that the source’s own summary requires. And the regional figure must reconcile against the wider grouping it sits within. Where a check fails, the failure is published in place of the table.
No cause is given for the Indonesian or the Philippine decline. No cause is published, and none is invented here.
No sector split by market, and no by-activity breakdown on 2025 data. The source of record for composition by activity is the ASEAN Secretariat's ASEAN Investment Report, whose latest edition is AIR 2025, on 2024 data. A 2024 composition is not carried forward and relabelled as a 2025 one; it is published under its own year. The by-activity tables return when AIR 2026 is published. The regional source-economy mix for 2025 is published separately by the Secretariat through ASEANstats, preliminary as at 1 July 2026, and appears in this edition as Table 2 on that basis alone.
No greenfield or project-finance figures for this region. Those are published only at the developing-Asia level, and presenting them as South-East Asian would be false precision.
No FDI stock figure. The annex row did not reconcile against an earlier published statement from the same body, so it is withheld pending a manual read rather than printed with a caveat.
The ten-year series and the stability classification are drawn from a second compilation, and are labelled as such. The UNCTAD annex table carries three reference years, not ten, so the decade series and the coefficient of variation beside it come from the World Bank’s World Development Indicators, which carries no 2025 value for Malaysia, Viet Nam, Myanmar or Lao PDR. Those four markets therefore show 2024 as their latest year, marked on the row. The two compilations are never added together and never compared line by line with the UNCTAD table.
Three balance-of-payments compilations of South-East Asian FDI are in circulation — UNCTAD's, the World Bank's World Development Indicators, and the ASEAN Secretariat's ASEAN FDI Database — and they do not agree market by market. This edition uses UNCTAD's throughout and mixes none of them. Three documented causes apply. Presentation basis: several members submit to the Secretariat on a directional basis while publishing their own balance of payments on the asset and liability principle — Singapore and Timor-Leste both state this in their metadata, and UNCTAD footnotes Brunei Darussalam, Viet Nam and the Philippines on its own table for the same reason. Manual vintage: Indonesia's submission to the Secretariat is made on the BPM5 standard by agreement, while Bank Indonesia's own balance of payments has followed BPM6 since 2014. Reporting vintage: Secretariat submissions are preliminary in the first instance and final typically two years after the reference period.
Coverage is the fourth reason, and it is why One source is used here. The World Bank's series publishes no 2025 value for Lao PDR, Malaysia, Myanmar or Viet Nam, and the ASEAN Secretariat's regional aggregates by activity exclude Timor-Leste. UNCTAD's South-East Asia grouping carries all eleven markets on one table for 2025, which is the coverage this publication requires.
No line-by-line reconciliation of these three series has been published by any of the three bodies, and none is asserted here.