Global Ocean Accounts Partnership Secretariat · GOAP Technical Guidance on Ocean Accounting · circular
This Circular specifies how a financial record is classified as ocean-related at the point that record is created — and what has to be true of that classification for the resulting aggregate to be reconcilable with the accounts. Upon completing it, readers will be able to design a blue tagging system for public expenditure that reconciles with government finance statistics, apply the same classification logic to sustainable finance instruments, reconcile the two resulting aggregates without double counting, publish a tagged figure with the metadata that makes it auditable, and join tagged expenditure to the ecosystem accounts covering the places that expenditure affected.
The Circular takes its ocean boundary from TG-3.3 Economic Activity Relevant to the Ocean and its functional classification from TG-1.1 OA and National Budget Processes, which sets out the COFOG classes carrying ocean-related government functions. It supplies the classification decision that populates the sustainable ocean finance aggregate measured in TG-2.6 Ocean-related Investment, and the eligibility and evidence rule on which the blue bond impact reporting in TG-1.7 OA and Multilateral Development Finance and the project appraisal in TG-1.8 OA and Project-Level Finance rest.
This Circular requires familiarity with:
Readers will also find the following useful, though they are not prerequisites: TG-2.6 Ocean-related Investment for the measurement of the aggregates a finance tag populates, TG-3.4 Flows from Economy to Environment for environmental protection expenditure, TG-3.7 Governance Accounts for the access-rights record that determines whether a coastal activity can be assessed at all, and TG-1.11 Ocean Accounts and Private Sector Decision-Making for the corporate disclosure frameworks that consume the same evidence.
Readers approaching Section 1 for the first time should begin with TG-1.0 Ocean Accounting and the Blue/Ocean Economy, the framing circular for the section, which sets out the ocean economy terminology used throughout, what an ocean account can and cannot certify about a blue economy claim, and the map from a policy question to the circular that answers it. It is recommended reading, not a prerequisite.
A tag is a classification attached to an individual financial record, applied at or near the point the record is created, and carried by that record through execution and reporting. A blue tag asserts that the record relates to the ocean, states in what respect, and retains the evidence and the identity of whoever made the assertion.
Two other methods already address the same question and neither is a tagging system. Governments and their partners have three options, set out in Table 1.
Table 1: Three methods for producing an ocean public expenditure figure
The three are complementary and the relationship is worth stating plainly, because a ministry will usually meet the first of them before it meets this Circular. The World Bank’s Blue Public Expenditure Review guidance note is the incumbent instrument for this question and it is well designed for what it does, which is to open a policy conversation. It sets no classification underneath its scope decision, and across its thirty pages it does not discuss tagging, COFOG, or the climate expenditure review methodologies from which tagging descends — natural capital accounts appear once, in a footnote. A Blue PER conducted on a budget that is already tagged is cheaper, repeatable, and can report a trend instead of a snapshot. Tagging is the standing infrastructure that a periodic review consumes.
Apportionment provides the external benchmark. Where the tagged envelope and the apportioned estimate diverge substantially and persistently, one of them is wrong, and the divergence is itself a quality signal (Section 3.8.2). A tagging system never compared against an apportioned estimate has no independent check on its own coverage.
Apportionment is also the cheapest thing to do first, and it is usually treated only as the check that comes last. Compiled before any tag is applied, the COFOG-based ocean share under TG-3.7 Section 3.3.5 states which functional classes carry material ocean-related expenditure and approximately how much, which is a map of where a tagging exercise should start and a defensible reason for leaving the remainder untagged in a first year. A ministry that apportions first knows the approximate size of the answer before it designs the system that will produce it, and can size the coding effort against the value at stake rather than against the number of budget lines. The same table then serves as the coherence check at Section 3.8.2 without further compilation.
Tagging systems for adjacent policy objectives are well established, and they are geographically spread: climate expenditure tagging in more than twenty countries across Asia, Africa and Latin America; biodiversity expenditure classification through BIOFIN’s Global Biodiversity Expenditure Taxonomy, applied in Malawi among others; SDG budget tagging, worked through in Colombia; and green budget tagging across OECD member states. Ocean-specific tagging is newer and thinner. UNDP lists green and blue public budget tagging among ocean finance instruments and names Blue Economy Budget Tagging in Indonesia as its example. The method of that system is not publicly documented at the time of writing, so this Circular records its existence without describing it.
What no other domain has is a compiled national account of the thing being tagged, which is what Section 3.1.3 turns on. The practical consequence for a compiler is that a blue tag will almost never be the first tag applied to a budget line. It will land alongside a climate tag, and increasingly a biodiversity tag, and Section 3.6 addresses what follows from that.
The reason a tag belongs in a guidance publication on ocean accounting is that a tagged budget line and an account entry can be the same object observed at two points in its life. An appropriation is made and tagged. It is executed, and the execution is recorded in the government’s financial system. In due course the expenditure appears in government finance statistics, in the environmental protection expenditure account where its purpose qualifies, and, where the accounts are compiled with sufficient granularity, against the ecosystem it was spent on.
Whether these are recognisably the same object depends on four choices made when the tag is designed. Each is cheap at design time and effectively irrecoverable afterwards:
A tagging system missing any of the four still produces a number, and that number will not join to anything.
· Figure 1.12.1
Source SEEA Central Framework (2012), Chapter 4 (environmental protection expenditure account, CEPA and CReMA classifications); IMF, Government Finance Statistics Manual 2014 (expenditure aggregates); TG-3.7, Section 3.3.5, Table 3.7.7 (COFOG mapping and the ocean-share apportionment method that supplies the benchmark).
The case for building a tagging system against the accounts, and not beside them, is that each supplies something the other cannot produce for itself. Table 2 sets out the exchange.
Table 2: The exchange between a tagging system and the ocean accounts
Three problems are solved by the pairing that neither solves alone. An ocean expenditure figure becomes available during budget formulation instead of two years after the fact. A published figure becomes auditable, because it reconciles to an aggregate compiled independently of the agency that produced it. And expenditure becomes joinable to measured ecosystem change, which is the precondition for ever asking whether ocean spending works.
This is also what distinguishes blue tagging from the climate, biodiversity and SDG systems it will sit alongside. Those have no compiled national account of the domain they tag against. Ocean accounting does, which means a blue tag can be checked against something, and theirs cannot.
A tag needs a boundary before it needs anything else, and the boundary belongs to TG-3.3 — a tagging system that defines its own has guaranteed that its output cannot be reconciled with the accounts. This is the principal difference from a Blue PER, whose scope is negotiated case by case and documented rather than classified.
What this Circular adds is the operational form of that boundary. An officer looking at a budget line applies three tests, and they do not always agree.
The activity test. Is the activity itself ocean-related, in the sense of TG-3.3? Four examples pass on activity alone — a fisheries licensing programme, a port dredging contract, a coastal ranger service, a marine research vessel.
The purpose test. Is the stated purpose of the expenditure the protection, restoration or sustainable management of a marine or coastal resource, in the sense of the CEPA and CReMA classes of the SEEA Central Framework? An upland reforestation programme whose stated objective is to reduce sediment load reaching a reef passes on purpose whilst failing on activity.
The place test. Does the expenditure occur in, or act on, the coastal zone or marine waters as delineated under TG-3.11 Sub-National Ocean Accounts? A wastewater treatment plant in a coastal city passes on place, and its ocean relevance is then a matter of degree.
Disagreement is the normal case, and a tagging system that states no ordering rule will produce inconsistent tags across agencies. This Circular recommends the following order, which places evidentiary defensibility above breadth.
A record satisfying the activity test is tagged in full. This is the expenditure that would appear in an ocean economy account on its own merits, and an auditor can confirm it without exercising judgement.
A record satisfying purpose but not activity is tagged, with the extent determined under Section 3.3.3 and the causal chain to a marine outcome documented. The upland reforestation example belongs here — tagging it is correct, tagging it in full is not, and tagging it without recording the sediment pathway makes the tag unreviewable.
The pathway is an account entry before it is a narrative. Residual flows from the economy to the environment are compiled under TG-3.4 Flows from Economy to Environment, and a sediment or nutrient flow recorded there from a catchment to a receiving marine unit is what the documented chain should cite. Where that flow account exists the causal chain is a reference; where it does not, the chain is an assertion by the tagging officer, and the metadata record should make clear which of the two it is.
A record satisfying place alone is tagged only where a marine benefit is identifiable and stated. Place is the weakest of the three tests and carries the greatest inflation risk, because in a coastal state a large share of all public activity occurs somewhere in the coastal zone. A system that admits place alone, without requiring a stated benefit, will tag a substantial part of the national budget and will be correct in a sense that serves nobody.
The stated benefit is what makes a place-only tag reviewable; the extent accounts are what make it testable. Such a tag asserts that expenditure in an administrative unit acts on the marine environment, and the ecosystem extent account for that unit records whether there is a marine or coastal asset there for it to act on. Where the crosswalk at Section 3.3.2 resolves the unit to a marine reporting unit with mapped extent under TG-3.1 Asset Accounts, the assertion has a referent an auditor can check without visiting the site; where it resolves to no marine unit at all, the tag rests on the officer’s judgement alone and should be refused. This is the control that turns the weakest of the three tests into one with a documented failure condition, and it is available only where extent accounts have been compiled.
The ordering is not arbitrary: it tracks how far each test can be carried into an account. An activity-test tag resolves to an ocean economy industry and can be reconciled against TG-3.3. A purpose-test tag resolves to a CEPA or CReMA class and can enter the environmental protection expenditure account. A place-test tag resolves only to a location, which is why it needs a stated marine benefit before it means anything. Evidentiary strength and account-readiness are the same ordering seen from two directions.
Where a record satisfies none of the three tests it is not tagged, and the assessment is recorded rather than left as a null, so that “assessed and not ocean-related” and “not yet assessed” remain distinguishable.
The attachment point in the chart of accounts determines everything the system can subsequently do, and it is the decision most often made by default.
At programme level. The tag attaches to a programme or sub-programme in the budget classification. This is the cheapest option — no change to transaction processing, and where most climate budget tagging sits. Its cost is that a programme is rarely wholly ocean-related, so proportional treatment does all the work, location is limited to whatever geography the programme reports, and the tagged figure cannot be traced to execution.
At activity or project level. The tag attaches one level below programme. Ocean relevance is far more often uniform at this level, so proportions are needed less often and are better founded when used. Most capital expenditure and most donor-funded activity already carries a project identifier, and a project usually has a location.
At transaction level. The tag attaches to individual commitments and payments in the treasury system. This gives execution data and full auditability, and it requires either a change to the financial management information system or a manual coding burden that in most settings will not be sustained.
The choice is also a choice about which accounts the tag can ever reach. A programme-level tag can populate a functional presentation and little else. An activity-level tag carries a purpose and a location, so it can reach the environmental protection expenditure account and, through the crosswalk, the ecosystem accounts. A transaction-level tag additionally distinguishes commitment from payment, which is what an account compiled on an accrual basis needs.
The recommendation is to attach at activity or project level and inherit upward to programme. That yields a tagged envelope reportable at any level of the budget classification whilst remaining traceable to an identifiable activity with a location. Systems beginning at programme level should treat migration to activity level as a planned second phase, not a later improvement, because location cannot be retrofitted to historical programme-level tags.
A blue tag is not a single flag. It carries four dimensions, each answering a different question and each mapping to something the accounts already hold. Table 3 sets them out.
Table 3: The four dimensions of a blue tag
Tying the function and purpose dimensions to established classifications rather than to a bespoke list is what makes the output comparable, and there is precedent for it in an adjacent domain: BIOFIN’s Global Biodiversity Expenditure Taxonomy classifies public biodiversity expenditure into nine primary categories with attribution rates tied to governmental functions.
The place dimension, and what is actually missing. A well-designed chart of accounts already carries a location code. Where a unified budget classification exists, it commonly includes a location segment reaching local-government level and coded to the national statistical geographic standard, and climate expenditure tags have been applied through such codes for around a decade. Compilers should not assume they need to build a spatial identifier; in many jurisdictions they need to use the one that exists.
The gap is elsewhere, and it is specific: a barangay code does not locate a reef. Administrative geography and ecosystem reporting units are different partitions of space. Marine and coastal accounting units follow habitat, depth, management area and fisheries management region — they neither nest within nor align to municipal boundaries. A great deal of ocean-relevant expenditure is coded to the coastal municipality from which a vessel, a ranger station or a project office operates, and not to the water it acts on.
The requirement is therefore a maintained crosswalk between the budget system’s location classification and the spatial units used by the ecosystem accounts, covering at minimum three cases — which administrative units front which marine reporting units, how expenditure coded to a land unit but acting seaward is allocated, and how expenditure coded to a national or headquarters location is treated. The crosswalk is standing infrastructure, not a one-off mapping, and where the same statistical office maintains both the geographic standard and the ecosystem accounts, it is cheap to build and nobody has yet built it.
The Philippines is the most fully documented case, and two features transfer. Climate Change Expenditure Tagging has run since 2015 under Joint Memorandum Circular 2015-01, and tags are applied through Unified Accounts Code Structure codes, whose Location segment reaches barangay level under the Philippine Standard Geographic Code maintained by the national statistical office. A compiler there does not need to build a spatial identifier. The same circular also decomposes rather than weights, which is the warrant for the preference order at Section 3.3.3.
The gap is the correspondence. The geographic standard locates a barangay; the ecosystem accounts report on habitat, depth stratum and management area; and nothing maps between them. Ocean accounting is what makes the gap visible and also what closes it, since the accounts define the marine reporting units the crosswalk has to reach. Country application belongs to the Section 5 case studies rather than to this Circular.
Where an activity genuinely has no location, such as a national fisheries policy unit or a marine science scholarship, the location is recorded explicitly as national. A null and a national are different statements, and conflating them breaks the join in Section 3.7.
· Figure 1.12.3
Source Illustrative of a budget classification carrying a location segment coded to a national statistical geographic standard: Government of the Philippines, Unified Accounts Code Structure, Location segment; Department of Budget and Management, Climate Change Commission and Department of the Interior and Local Government, Joint Memorandum Circular 2015-01, 24 March 2015, under which tags are applied through those codes.
Most tagged budget lines are not wholly ocean-related. Three methods are available and they are not equivalent. They form a preference order, and a compiler should exhaust each before moving to the next.
First, decompose. Where the budget structure permits, identify the components of the programme, activity or project that are ocean-related and tag those components in full, leaving the remainder untagged. This is what established climate expenditure tagging instruments direct: where addressing the objective is the main purpose, or one of the main purposes, of a programme, activity or project, the entire budget of that item is tagged, and otherwise the agency identifies only the components that directly address the objective and includes only those components’ expenditure. Decomposition has a decisive property — it partitions an appropriation into amounts that remain amounts, so the result is additive, auditable against the budget, and free of any judgement about proportion.
Decomposition also has an accounting payoff that weighting destroys. Components identified this way can be classified straight to CEPA and CReMA classes and carried into the environmental protection expenditure account, because each is a whole amount with a single purpose. A coefficient applied across a mixed line produces a figure that belongs to no purpose class at all, and so cannot enter the account. Where a compiler decomposes once against the account structure, the decomposition is reusable in every subsequent year.
Second, measure a share. Where the line cannot be decomposed, derive a proportion from an operational quantity — the share of shoreline length treated, the share of vessel days in marine waters, the share of a research programme’s funded projects with a marine subject. Record the derivation. A measured share is defensible in a way a convention is not, and it should be labelled as measured in the metadata record.
The quantities that make a share measurable are, in most cases, already compiled. A spatial proportion is derived from the areas and shoreline lengths in the extent accounts of TG-3.1 Asset Accounts; an activity proportion from the gross catch and extraction quantities in the physical flow accounts of TG-3.2 Flows from Environment to Economy, or from the spatially mapped vessel-days compiled under TG-4.10 Extending Fisheries Statistics; and an industry proportion from the output and employment aggregates of TG-3.3.
A compiler holding these tables derives a proportion where a compiler without them must assign a coefficient, and that is the most direct contribution the accounts make to a tagging system: account coverage determines how much of the tagged envelope rests on measurement and how much rests on convention. It follows that the composition of the envelope by extent method, published under Section 3.3.4, is a coverage statistic for the accounts as much as a quality statistic for the tag, and a fall in the coefficient-derived share is the return on each additional account compiled.
Third, assign a coefficient. Only where neither of the above is possible, apply a documented ordinal scale. This is the method of the OECD DAC Rio markers, and the two properties a compiler should know about it are that the coefficients are not standardised — applied rates for the “significant” marker range across providers from zero to 100 per cent, with no common reporting standard — and that the marker system itself describes its output as “descriptive rather than strictly quantitative”, allowing only “an approximate quantification”.
An ordinal scale mapped to percentages is a convention adopted for tractability. This is visible in the adjacent domain as well: BIOFIN published its biodiversity expenditure taxonomy in 2024 with the attribution rates themselves marked for a future version. A compiler using coefficients should record that they are adopted as a convention, and should not present a coefficient-weighted figure as an estimate of ocean expenditure.
The two-figure rule, and where it applies. A figure produced by coefficients is not additive with any fiscal aggregate. The amounts appropriated are whole amounts, and applying a 50 per cent coefficient to a line does not create a 50 per cent appropriation, so a sum of coefficient-weighted lines corresponds to nothing in the budget, the treasury ledger or government finance statistics. Where any coefficient has been applied, two figures are published:
Publishing the weighted estimate alone is the most common defect in existing tagging systems, and it produces a figure that no auditor can confirm and no analyst can decompose. Where a system has reached full decomposition, the two figures converge and the rule costs nothing, which is a further reason to prefer decomposition.
· Figure 1.12.2
Source Department of Budget and Management, Climate Change Commission and Department of the Interior and Local Government (Philippines), Joint Memorandum Circular 2015-01, Revised Guidelines for Climate Change Expenditure Tagging, 24 March 2015, Annex A, Step 2 (decomposition rule); OECD Development Assistance Committee, Rio Markers for Climate: Handbook (marker scale, and the description of markers as descriptive rather than strictly quantitative); OECD, survey of coefficients applied to Rio-marked flows (range across providers).
Three roles must be distinct.
The tagger is normally the spending agency, the only party with knowledge of the activity. Tagging by a central agency from budget documents alone produces weaker tags, because the documents rarely contain what the purpose and place dimensions need.
The rule-setter is the finance or planning ministry, which issues the tagging guidance, the boundary definition, the classification lists and the crosswalk. Where agencies set their own rules the resulting aggregate is not comparable across agencies or across years.
The verifier is independent of the spending agency. Its function is to sample, to test tags against the documented rule, and to publish an error rate. A system that publishes no error rate asserts a precision it has not established. Established tagging instruments attach a formal quality assurance review to the tagging cycle, which is a workable model.
Tag inflation is the characteristic pathology, and it is a rational response to incentives. Where tagged expenditure attracts political credit, counts toward an international commitment, or improves access to concessional finance, agencies will tag generously, and the inflation will appear in the tests that admit the most judgement. Four controls are available and none is sufficient alone:
The last of these is why the accounts belong in the institutional design and not only in the methodology. A tagging system whose only reference point is itself will drift, and the environmental protection expenditure account and the COFOG-based ocean share are compiled by a statistical office with no interest in the tagged total. Assigning the annual comparison to that office, rather than to the finance ministry or the spending agency, gives the arrangement an external anchor that no amount of internal sampling supplies.
The second application of the same operation classifies financial instruments and the activities they finance. The classification decision is made by a lender, an issuer, a verifier or a regulator, and the tests it applies are set by a sustainable finance taxonomy where the jurisdiction has one.
A taxonomy performs two functions: definition, which activities are eligible, on what tier, subject to which do-no-significant-harm tests and exclusions; and disclosure, what the financed entity must report about the environmental characteristics of what it does. Each is a measurement problem before it is a regulatory one.
Taxonomies with a marine sector are recent, few and unaligned. The ASEAN Taxonomy for Sustainable Finance completed technical screening criteria across the focus and enabling sectors of its Plus Standard in its fourth version, with a multi-tier structure framed to accommodate the differing development stages of member states. Indonesia’s national taxonomy acquired technical screening criteria for fisheries and marine activities in its third version, classifying an activity as green, as transitional, or as neither, and applying the full criteria only to the corporate non-MSME segment whilst offering a simplified decision tree to the remainder. Coverage elsewhere is partial: the European Union taxonomy carries an environmental objective on the sustainable use and protection of water and marine resources, and market guidance exists for bonds financing the sustainable blue economy and in the Climate Bonds Initiative marine criteria. No international blue finance taxonomy exists.
One methodological consequence matters more than the rest. A tiered criterion is more data-intensive than a binary one — a binary test asks whether an activity appears on a list, whilst a tiered test asks where on a gradient a particular asset sits, which requires the asset to be measured, spatially explicitly, by a recognised method, and repeatedly. A criterion expressed as improvement, or as a threshold sustained over time, cannot be evaluated against a single survey. It requires a time series, which is what an account compiled on a cycle is and what a commissioned baseline study is not.
The compilation task. A compiler serving taxonomy users should build and maintain a correspondence table with one row per applicable screening criterion, recording: the account table that supplies the evidence; whether a snapshot exists for it; and whether a repeat-observation series exists for it. On the asset side, resource stocks map to the SEEA Central Framework categories and ecosystem assets to the reference typology used in TG-3.1 Asset Accounts, and any divergence between a national ecosystem typology and the reference classification is documented explicitly, since the comparability of resulting disclosures depends on it. On the disclosure side, extent accounts with protected-area overlays serve location-of-assets disclosure, condition accounts serve dependency and impact assessment, physical flow accounts serve pressure metrics, and investment indicators serve nature-related finance metrics; the route by which a reporting entity consumes national accounts is set out in TG-1.11 Section 3.2.
The distance between the two ratings in that table — criteria with a snapshot against criteria with a series — is the finding most worth publishing, because it states exactly which criteria the jurisdiction can currently evidence and which it cannot.
Two instrument structures carry a blue tag and they mean different things.
A use-of-proceeds instrument commits the issuer to spending the proceeds on a defined list of eligible activities. The tag attaches to the proceeds, the reporting obligation is an allocation report, and the evidence required is an eligibility determination for each financed activity, made at issuance.
A performance-linked instrument commits the issuer to a measured outcome, with a financial consequence if the outcome is not met. The tag attaches to performance against a key performance indicator, and the proceeds are unrestricted. The evidence required is a baseline, a target, and an independently produced series capable of showing whether the target was met.
Aggregating the two destroys the distinction that a user of the figure most needs. A blue finance total mixing them states neither how much money was directed to marine activity nor how much marine outcome was contracted for. Use-of-proceeds and performance-linked instruments are reported separately.
This is the point at which the accounts stop being a reference and become the instrument. An account table compiled on a cycle is well suited to serve as the performance metric in a performance-linked instrument, because it is independently produced, independently verifiable, and outside the issuer’s control. That is a stronger position than the bespoke monitoring such instruments ordinarily rely on, and it is available only where compilation is recurrent (Section 3.9).
Most jurisdictions have no applicable marine taxonomy, and comparable blue finance figures are still needed. The three-tier convention in Table 4, adopted from GOAP Secretariat material prepared for national planning use, allows publication without waiting for taxonomy harmonisation.
Table 4: Interim blue finance classification convention
Every figure published under this convention carries a metadata record covering instrument type, issuer and sector, jurisdiction, taxonomy or criteria applied, verification status, whether the instrument is use-of-proceeds or performance-linked, the reporting period, and whether the figure is a stock or a flow, consistent with TG-2.6.
The third tier is where ocean accounting does the most work. Where no taxonomy applies, the compiler has to document the criteria applied, and an account table is the most defensible thing to document them against: extent and condition for the asset financed, and a compilation date establishing that the evidence is current. A criteria-documented figure anchored to a published national account is a materially stronger disclosure than one anchored to a consultant’s baseline study, and it costs less at the second use.
The tiers are reported separately and never summed into a single blue finance figure. A self-labelled total and a verified total answer different questions, and a reader who cannot see the split cannot tell which question the number answers.
The tagged public expenditure envelope of Section 3.3 and the blue finance total of Section 3.4 are drawn from different populations and answer different questions. The first is an expenditure view — what was spent, and on what. The second is a financing view — where money came from, and on what terms. They overlap, and the overlap is largest in exactly the jurisdictions most likely to publish both.
Three overlaps recur.
Sovereign issuance. Where a government issues a blue bond, the proceeds are appropriated and executed as public expenditure. The same money appears once in the blue finance total as an instrument and again in the tagged envelope as expenditure. It is one flow counted twice.
On-lending and public financial institutions. Where a national development bank lends to a private borrower using sovereign-guaranteed funds or an on-lent multilateral facility, the same underlying capital can be counted at the multilateral disbursement, at the sovereign appropriation, and at the sub-loan.
Budget support and grants. External finance recorded as an ocean-related grant by the provider is, in the recipient’s accounts, an appropriation, and where the recipient tags it, it appears in both.
The rule is simple and is violated routinely: the two aggregates are never summed. A country wishing to publish a single ocean finance figure must publish a bridge, and the bridge has an elimination line. Table 5 gives the minimum structure.
Table 5: Bridge between the financing view and the expenditure view
The bridge is built on the unweighted envelope, for the reason given at Section 3.3.3. A bridge built on coefficient-weighted figures reconciles nothing, because neither side corresponds to a payable amount.
Both sides of the bridge have an account counterpart, which is what keeps it from being an arbitrary reconciliation. The expenditure view resolves to the environmental protection expenditure account and the COFOG-based functional presentation. The financing view resolves to the supplementary “of which” classification of financial instruments measured in TG-2.6. The elimination at line B is therefore not a plug; it is the amount that appears in both account presentations, and a compiler who can locate it in each has demonstrated that the bridge closes.
Where a compiler cannot identify line B, the bridge cannot be constructed, and the two aggregates are published separately with a note stating that they overlap by an unquantified amount. That is an honest presentation, and a single summed figure is not.
Where a government already runs climate expenditure tagging, and increasingly biodiversity tagging alongside it, a blue tag lands on records that are already tagged. That overlap is the expected result, and the only error available is to treat it as a conflict.
Tag dimensions are orthogonal, not a partition. A mangrove rehabilitation appropriation is legitimately climate-relevant on both mitigation and adaptation grounds, and biodiversity-relevant, and blue-relevant. Forcing a single classification would require the compiler to discard true information.
This is settled practice in the oldest tagging system in existence, and this Circular applies it rather than proposing it. The OECD DAC Rio marker handbook is explicit: activities may qualify for more than one marker, and to avoid double or triple counting the same activity, “aggregate figures for biodiversity, climate change mitigation, climate change adaptation and desertification-related development finance should not be added up”, with statistical presentations either prepared for one marker at a time or presenting the overlap explicitly.
Two requirements follow for a blue tag.
Tag aggregates are never summed across dimensions. Adding a climate total to a biodiversity total to a blue total counts the mangrove line three times and has no interpretation.
Intersections are published. A blue tagging system operating alongside others reports a cross-tabulation of the tagged envelope, at minimum distinguishing blue-only, blue and climate, blue and biodiversity, and all three. The intersection cells usually carry more policy content than any single-tag total, because they identify the expenditure serving several national commitments at once, which is the expenditure a finance ministry is most interested in protecting through a consolidation.
The intersection table is itself an account presentation, and should be built as one. Combined presentations in TG-3.8 exist precisely to set several classifications against a common population without summing them, and a blue-climate-biodiversity cross-tabulation is that structure applied to appropriations. Building it as a combined presentation rather than as a spreadsheet has a practical consequence: the intersection can then be carried into the accounts alongside the expenditure it describes.
Where the same record carries different extents in different tagging systems, the discrepancy is disclosed and not reconciled. The systems answer different questions, and there is no reason the shares should sum to anything in particular.
This section is the payoff of the crosswalk requirement at Section 3.3.2.
The join is between two tables. On one side, tagged expenditure aggregated by spatial unit, purpose class and period. On the other, the ecosystem extent and condition accounts for the same units and periods, compiled per TG-3.1 Asset Accounts and the relevant Section 6 thematic circulars. Four conditions make it possible, and all four are design choices:
The join supports description. It states that a given quantum of blue-tagged expenditure, of a given purpose composition, occurred in a spatial unit over a period, and that the unit’s ecosystem extent and condition changed by a measured amount over a related period. That is a coincidence of two measured series, and it is worth publishing on its own.
The join does not support attribution. It does not establish that the expenditure caused the change, and a compiler presenting it as though it did will be wrong in the common case, because condition changes are driven overwhelmingly by pressures the expenditure did not address. Attribution requires a counterfactual, and the designs available — control sites, before-after-control-impact, matching, regression discontinuity — are set out in TG-1.3 Section 3.4.3 and applied to intervention accounting in TG-6.15 Ocean Accounting for Nature-based Solutions. Tagging supplies the expenditure series those designs need, and substitutes for none of them.
A joined table most reliably identifies spatial units where ecosystem condition is declining and blue-tagged expenditure is absent or negligible, and the converse, units carrying substantial expenditure with no condition series against which anything can be assessed.
Neither finding requires a causal claim, both are robust to the attribution limit above, and both are directly actionable: the first is a budget allocation question, and the second is a measurement programme question. The second is also a commissioning instrument for the accounts themselves. A list of spatial units carrying material blue-tagged expenditure and no condition series is a costed, prioritised case for extending ecosystem account coverage, addressed to the ministry already spending money there. A compiler producing nothing else from this Circular should produce these two lists.
A tag without metadata is an assertion. Table 6 specifies the minimum record retained for every tagged line.
Table 6: Minimum tag metadata record
Each field exists because an account consuming the tag needs it. Purpose class and location are what the EPEA and the ecosystem accounts read directly. Extent method and relevance basis are what allow an account compiler to decide whether a tagged amount may enter an aggregate at all, since a coefficient-derived figure may not. Verification status and tagger identity carry the provenance that TG-0.7 requires of any input to a published account.
The rule version field is most often omitted and is what makes a time series usable. Tagging rules are revised, and a tagged envelope compiled under a 2027 rule is not comparable with one compiled under a 2025 rule unless the change is recorded and, where material, the earlier series restated.
The quality dimensions in TG-0.7 Quality Assurance Principles apply, and three of them bite particularly.
Accuracy is assessed by sampling. The verifier draws a sample of tagged lines, re-applies the published rule independently, and publishes an error rate with its confidence interval. Errors are reported in both directions — incorrectly tagged, and incorrectly untagged. A system sampling only tagged lines will never detect under-coverage, which in a first year of operation is usually the larger error.
Comparability over time is protected by the rule version field and by restatement when rules change materially. Its main threat is budget classification reform, which is common and which severs the link between a historical tag and any current budget line. A crosswalk between successive budget classifications should be maintained deliberately, alongside the spatial crosswalk at Section 3.3.2.
Coherence is assessed against the apportioned estimate compiled under TG-3.7 Section 3.3.5, annually, and against the environmental protection expenditure account for the purpose-classified subset. This is the external check, and the only quality assurance step whose result does not depend on the tagging system’s own procedures.
The requirements above are not all attainable at once. Three tiers describe what a country can produce given what it has, following the compilation pathway convention used elsewhere in this publication. The tiers describe capability, not ambition: a country at the foundational tier should design for the advanced tier from the start, because the four choices at Section 3.1.2 are cheap now and unrecoverable later.
Foundational. Tags attach at programme level, with the function and purpose dimensions and whatever extent method the budget structure supports. The tagged envelope is published, with its composition by boundary test and extent method, and the weighted estimate alongside it wherever coefficients have been used. No crosswalk, therefore no join to the accounts. The output is a defensible expenditure figure and a coverage baseline. It requires the tagging guidance itself, a coding exercise during formulation, and the willingness to publish two figures.
Developing. Tags attach at activity or project level. The crosswalk is built for capital works, protected area management and area-based programmes, which will typically cover most of the envelope by value. Verification sampling begins and an error rate is published. The bridge to the financing view is attempted where sovereign issuance exists. The output additionally supports the gap analysis at Section 3.7.2 for the covered subset. It requires a change to the budget classification or the activity register, and a verification function with a mandate.
Advanced. Tags attach at activity level with execution data from the treasury system and carry the full metadata record at Table 6. Crosswalk coverage is complete, including explicit national coding. Cumulative expenditure series by spatial unit are maintained. Tag intersections with climate and biodiversity systems are published. The joined table at Section 3.7 is produced annually alongside the accounts. It requires a financial management information system able to carry the tag through execution, and recurrent compilation of the ecosystem accounts on the other side of the join.
The binding constraint at every tier is the second item on that last list. A tagging system can be built in one budget cycle; an account it can be joined to requires compilation on a cycle, and the GOAP Secretariat’s observation from country programmes is that the difficult step for a national accounting initiative is the second compilation and not the first. Where the ecosystem accounts are compiled once as a pilot, the advanced tier is unattainable however good the tagging system is, and the tagging system will in time be judged on outputs it was never able to produce.
Tagging measures intent, not effect. A tagged appropriation records what a government decided to fund and for what stated purpose. It records nothing about whether the money was well spent, whether the activity was completed, or whether anything in the marine environment changed. Section 3.7 sets out how far the accounts take a compiler beyond intent, and it is not far without a counterfactual.
Coefficients have no empirical foundation. Where Section 3.3.3’s third method is used, the resulting figures should not be described as estimates of ocean expenditure, and should not be compared across jurisdictions applying different scales. Applied Rio marker coefficients already range from zero to 100 per cent across providers with no common standard.
Coverage is bounded by the budget. Blue tagging observes appropriated public expenditure. It does not observe extrabudgetary funds, state-owned enterprise expenditure outside the budget, unconsolidated sub-national own-source expenditure, or in-kind contributions. In several ocean-relevant domains, notably port authorities and fisheries agencies operating on retained revenue, the unobserved share may be substantial and should be estimated and disclosed.
The finance side rests on jurisdictional taxonomies that do not align. A blue finance figure classified under one national taxonomy is not comparable with one classified under another, and the interim convention at Section 3.4.3 makes the incomparability visible without resolving it. No international blue finance taxonomy exists, and this Circular does not propose one.
Tagging systems are political instruments. Where the tagged figure carries consequences it will be gamed, and the controls at Section 3.3.4 reduce that without eliminating it. A tagged envelope that grows rapidly without a corresponding change in the budget should be investigated as a tagging phenomenon before it is reported as a policy achievement.
The Circular is badged Emerging. No adopted international standard governs the tagging of expenditure or finance as ocean-related, and ocean-specific practice is thin (Section 3.1.1). The classifications relied on here — COFOG, CEPA, CReMA, the SEEA structures — are standards; the way they are assembled into a tag is not, and it should be expected to change as practice develops.
This Circular has been approved for public circulation and comment by the GOAP Technical Experts Group in accordance with the Circular Publication Procedure.
Authors: [To be confirmed]
Reviewers: [To be confirmed]
Cite this report
OA and Blue Tagging of Public Expenditure and Finance (2026) (GOAP Technical Guidance on Ocean Accounting, TG-1.12, v0.1.0). GOAP Secretariat. https://guidance.oceanaccounts.org/circulars/section-1/1-12-blue-tagging/
guidance.oceanaccounts.org/circulars/section-1/1-12-blue-tagging
© 2026 Global Ocean Accounts Partnership Secretariat · CC BY-NC 4.0 · The report, its PDF and this deck are built from the same source.
Swipe on touch screens. Print gives one landscape page per slide. The address keeps the slide number, so a link opens on the right slide. On a phone in portrait the deck becomes a scrolling document.