Introduction

The Federated States of Micronesia occupies an unusual position in this series: most of the fiscal-sustainability questions other Pacific papers in this project address as open risks are, for FSM, already substantially resolved. Public debt fell to 8.9% of GDP in FY2025. Trust-fund assets reached 510% of GDP, having grown from 323% just two Article IV cycles earlier. The 2023 renewal of the Compact of Free Association with the United States materially improved the country's medium-term debt dynamics. Growth has been positive, if modest and volatile (2.3% in FY2024, moderating to 1.1% in FY2025 and a projected 1% in FY2026), driven substantially by the fisheries sector and Compact-funded public investment.

What makes FSM's fiscal position genuinely interesting for this paper is not fragility but structure. FSM is a federation of four states — Chuuk, Kosrae, Pohnpei and Yap — each with its own government, and fishing-licence revenue, one of the country's most important domestic income sources, is shared between the national government and the states under a revenue-sharing arrangement that the IMF's own 2026 mission statement flagged directly: states increased expenditure in response to a higher revenue envelope from a greater share of fishing fees, a fiscal-coordination dynamic with no precedent in the more centralised administrations covered elsewhere in this series. Separately, FSM's own 2019 record shows a specific instance of fiscal discipline worth generalising: the authorities' decision to save most of a corporate-tax windfall into the FSM Trust Fund rather than spend it, a decision the IMF welcomed explicitly at the time but which has not, to this paper's knowledge, been formalised into a standing rule that would apply automatically to the next windfall.

Compact sector grants were scheduled to decline after FY2023 under the original Compact timeline, a transition the 2023 Compact renewal has since eased but not eliminated as a medium-term planning question; the country's stated priority, in the IMF's own words, is continued tax reform, targeted expenditure discipline, and structural reform to reduce aid dependency and support private-sector growth, an agenda IMF missions have been recommending in broadly similar language for well over a decade.

The research question this paper takes up has three parts: how the national-state fishing-fee revenue-sharing formula should be modelled so that state-level fiscal response to a shared, volatile revenue source is coordinated rather than left to each state's independent judgement; how a trust fund already at 510% of GDP should be understood going forward, given that its scale changes the nature of the sustainability question from whether it will be adequate to how its growth and eventual use should be governed; and how FSM's own demonstrated capacity to save a windfall, shown once in 2019, should be built into a standing rule rather than relying on the same discretionary judgement being exercised well every time. As throughout this series, the organising identity for the tax-side question splits the gap between potential and actual corporate and gross-receipts tax revenue into a policy component and a compliance component:

R^{pot}_{GRT+CIT,t} - R_{GRT+CIT,t} = \underbrace{\left(R^{pot}_{GRT+CIT,t} - R^{law}_{GRT+CIT,t}\right)}_{\text{policy gap: gross-receipts design, still-pending reform}} + \underbrace{\left(R^{law}_{GRT+CIT,t} - R_{GRT+CIT,t}\right)}_{\text{compliance gap}}

Current tax issues

2.1 A four-state federation sharing a volatile revenue source

FSM's national government and its four constituent states each administer their own budgets, and fishing-licence revenue is shared between them under an arrangement that has become materially more consequential as fishing-fee income has grown. The IMF's 2026 language is specific: a greater share of fishing fees flowing to the states led directly to increased state-level expenditure. Fishing-fee revenue is itself volatile, tied to tuna markets and licensing negotiations under the regional Vessel Day Scheme; a revenue-sharing formula that passes that volatility through to four independent state budgets without any coordinating mechanism risks four separate pro-cyclical spending responses to a single shared shock.

2.2 A trust fund whose scale changes the question

At 510% of GDP, FSM's combined trust-fund assets are large enough that the conventional adequacy question this series has asked for Kiribati's RERF, Marshall Islands' CTF and Nauru's NTF — will the fund be big enough — is not FSM's binding constraint. The more relevant questions are about governance: how fast should the fund continue growing, what draws (if any) are appropriate as Compact grants evolve, and whether a fund of this scale relative to the domestic economy creates its own risks (for instance, to monetary and financial stability, given FSM's use of the US dollar) that a smaller fund would not.

2.3 A demonstrated but informal windfall-saving practice

The 2019 decision to save most of a corporate-tax windfall into the FSM Trust Fund, welcomed explicitly by the IMF at the time, shows FSM's authorities are capable of the fiscal discipline a windfall requires. What is not evident from the public record is whether that decision reflected a standing policy or a one-off judgement call specific to that year's circumstances; this paper treats the latter as the more cautious assumption and proposes formalising the practice accordingly.

2.4 Compact grant decline, eased but not eliminated

The original Compact funding schedule anticipated declining sector grants after FY2023; the 2023 Compact renewal has substantially improved the medium-term picture, but the underlying transition from grant-funded to trust-fund-and-domestic-revenue-funded spending remains a live medium-term planning question, not one the renewal fully resolved.

2.5 Tax reform recommended for over a decade, still in progress

IMF mission language recommending continued tax reform in FSM dates back to at least the early 2010s and recurs through the 2026 Article IV consultation. The gross-receipts tax (GRT) that forms part of FSM's tax base is the specific instrument most often cited as needing reform, though the pace of that reform has been slow enough that recommending it has itself become a recurring feature of FSM's Article IV cycle.

2.6 Emigration and a shrinking domestic base

FSM citizens have free-migration rights to the United States under the Compact, and continued outward emigration has been a recurring IMF concern since at least 2010, narrowing the domestic tax and labour base even as the trust fund grows. This tension, a shrinking population alongside a growing fund, is a genuinely distinctive feature of FSM's fiscal position within this series.

Models

3.1 Model A: national-state fishing-fee revenue-sharing and fiscal coordination

Model A represents the four states' fiscal response to shared fishing-fee revenue explicitly, rather than leaving each state's expenditure decision implicit in its own separately reported budget:

FF_t = \text{TotalFishingFeeRevenue}_t, \qquad FF^{state}_{s,t} = \phi_s \cdot FF_t, \qquad \sum_s \phi_s = 1

where phi_s is state s's share under the current formula. A coordination indicator flags where states' aggregate expenditure response to a fishing-fee windfall exceeds what a smoothed, multi-year view of that revenue would support:

\text{ExpenditureResponse}_{s,t} = \frac{\Delta \text{StateExpenditure}_{s,t}}{\Delta FF^{state}_{s,t}}, \qquad \text{flag if } \text{ExpenditureResponse}_{s,t} > 1

a ratio above 1 meaning a state increased spending by more than its share of the windfall itself, implying it drew on other reserves or borrowing to expand further, a pro-cyclical amplification of the shared shock rather than a dampening of it. A jointly agreed smoothing mechanism, structurally analogous to Kiribati's trend/cycle decomposition (Paper 8) but applied at the inter-governmental level, would base each state's budgeted fishing-fee allocation on a trailing multi-year average rather than the current year's collection, with above-average years building a shared or state-level reserve against below-average ones.

3.2 Model B: Trust Fund growth-trajectory and glide-path governance

Model B treats FSM's trust-fund assets as large enough that the governing question is trajectory and use, not adequacy:

TF_{t+1} = TF_t(1+r) + C_t - D_t, \qquad \frac{TF_t}{GDP_t} \text{ tracked explicitly over time}

where C_t is contributions (including any windfall quarantined by Model C) and D_t any draws. Rather than testing whether TF_t reaches a target, Model B tests the trajectory of TF_t/GDP_t against reasonable governance bounds — for instance, whether continued growth at anything like the recent pace (323% to 510% of GDP in two Article IV cycles) is intended policy or an unexamined by-product of Compact-era fiscal surpluses, and what glide path toward Compact-grant decline the fund's own growth should be matched against:

\text{GlideGap}_t = \text{ProjectedComptactGrantDecline}_t - \text{PlannedTrustFundDrawCapacity}_t

A positive GlideGap indicates the fund's planned draw capacity would not fully replace declining Compact grants at the projected pace, even at FSM's exceptional fund scale; a substantially negative one suggests the fund could support materially more domestic spending than current policy draws on, a genuinely different diagnostic than any prior country paper in this series has needed to make.

3.3 Model C: windfall detection and quarantine rule

Model C formalises FSM's 2019 precedent into a standing, rules-based mechanism rather than case-by-case discretion. A windfall is defined relative to a trailing trend, using the same decomposition logic as Kiribati's Model C (Paper 8):

Revenue_t = Trend_t + Windfall_t, \qquad Windfall_t = Revenue_t - Trend_t \text{ where } Windfall_t > \tau

for a materiality threshold tau. The quarantine rule then directs a fixed share of any detected windfall automatically to the Trust Fund, rather than requiring a fresh discretionary decision each time:

\text{QuarantinedAmount}_t = \psi \cdot \max(0, Windfall_t), \qquad \psi \in (0,1]

applied uniformly across corporate tax, GRT and fishing-fee revenue sources, so the 2019 corporate-tax precedent becomes a general policy rather than a source-specific one-off.

Algorithms and illustrative results

4.1 Algorithm 1: fishing-fee revenue-sharing coordination monitor

ANNUALLY:
  FF_t <- total fishing licence fee revenue collected nationally
  FOR each state s:
    FF_state_s <- phi_s * FF_t   # per the current sharing formula
    delta_expenditure_s <- StateExpenditure_s,t - StateExpenditure_s,t-1
    response_ratio_s <- delta_expenditure_s / (FF_state_s,t - FF_state_s,t-1)
    IF response_ratio_s > 1: FLAG state s for pro-cyclical amplification
  trend_FF <- trailing 5-year average of FF_t
  FOR each state s:
    smoothed_allocation_s <- phi_s * trend_FF
    IF FF_state_s,t > smoothed_allocation_s:
      surplus_s <- FF_state_s,t - smoothed_allocation_s
      RECOMMEND surplus_s contributed to a state or shared reserve
PUBLISH the coordination summary to the national-state Joint Economic
  Management forum for discussion ahead of each budget cycle

4.2 Algorithm 2: windfall detection and quarantine

QUARTERLY, for each major revenue source (corporate tax, GRT, fishing fees):
  FIT trailing trend (e.g. 5-year moving average or a simple state-space trend)
  windfall_t <- max(0, actual_t - trend_t)
  IF windfall_t > materiality_threshold:
    quarantined_amount <- psi * windfall_t
    TRANSFER quarantined_amount to the FSM Trust Fund automatically
    REPORT the transfer and its basis to the national legislature
  remaining_amount <- actual_t - quarantined_amount
  AVAILABLE for ordinary budget use

4.3 Illustrative results: Trust Fund trajectory at exceptional scale

Using illustrative inputs (nominal GDP on the order of USD 450 million, a starting trust-fund value consistent with the reported 510% of GDP, a 5% real return, and roughly 1.5% real GDP growth), continuing to accumulate at anything close to the pace implied by the recent 323%-to-510% shift (an average of roughly 62 percentage points of GDP per year) would push the fund toward approximately 1,445% of GDP within a decade — a figure large enough that it should itself prompt a governance conversation about intended policy versus unexamined momentum. Even a materially slower future accumulation pace, illustratively a fifth of the recent rate, still leaves the fund at roughly 862% of GDP after a decade. Both scenarios support this paper's central claim: FSM's trust-fund question, unlike every other sovereign-fund case in this series, is not whether the fund will be adequate, but how its continued growth and eventual use should be governed.

4.4 Illustrative arithmetic: the fishing-fee coordination gap

If a state's share of a fishing-fee windfall is, illustratively, 15% of the national total, and that state's expenditure increase in the same year exceeds 15% of the total windfall value, Algorithm 1's response-ratio flag activates — a signal that the state financed additional spending beyond its direct share, whether through drawing down its own reserves or through borrowing, amplifying rather than smoothing the shared national windfall at the state level. This is exactly the dynamic the IMF's 2026 language describes qualitatively (states increased expenditure in response to a higher revenue envelope); Algorithm 1 gives the Joint Economic Management forum a concrete, state-by-state figure to discuss rather than a general observation.

Solutions, software, data science and AI models

5.1 Issue-to-solution map

5.2 Model catalogue

5.3 Reference architecture

flowchart TD
  A[Sources: national and state budget records, fishing licence fee collections, FSM Trust Fund and Compact Trust Fund statements, GRT/CIT returns] --> B[Ingestion and data quality]
  B --> C[Warehouse: bitemporal, national-state linked]
  C --> D[Fishing-fee coordination monitor]
  D --> E[Joint Economic Management forum dashboard]
  C --> F[Windfall detection service]
  F --> G[Automated Trust Fund transfer]
  C --> H[Trust Fund trajectory and glide-path model]
  H --> I[National government fiscal dashboard]

The coordination monitor and windfall detection service both run on an annual/quarterly cadence tied to the budget process; the trajectory model runs on demand for medium-term fiscal planning.

5.4 Software stack

5.5 Governance of AI use

• Coordination flags inform, not override, state fiscal autonomy. Model A's response-ratio flag is a discussion input for the Joint Economic Management forum; it does not constrain any state's constitutional budget authority.
• The quarantine rule is legislated, not discretionary once adopted. Once psi and the materiality threshold are set through the national legislature, Model C's transfers should be automatic, precisely to remove the year-by-year discretion this paper argues is FSM's current gap.
• Trust Fund trajectory scenarios are published in full, not selectively. Given the fund's unusual scale, this paper recommends publishing both the continued-pace and slower-pace scenarios in section 4.3 together, so policy discussion is not anchored to only the more dramatic figure.
• No model output determines Compact-related negotiating positions. These models inform FSM's own domestic fiscal planning; they are not positioned as inputs to US-FSM Compact negotiations, which remain a bilateral policy matter outside this paper's scope.

5.6 Team and indicative effort

Given FSM's federated structure, a small national-level analytics function — one lead working with the Department of Finance and Administration, plus a designated liaison in each of the four states for Model A specifically — is proportionate. The Trust Fund trajectory model (Model B) is primarily a national-government and Trust Fund administrator exercise. Cost and staffing figures are estimates.

Limitations

• The Trust Fund trajectory figures in section 4.3 use illustrative GDP and fund-value inputs consistent with the order of magnitude in published IMF figures, not FSM's or the Trust Fund administrators' actual detailed financial statements, which this paper does not have access to.
• The fishing-fee revenue-sharing formula (the phi_s values in Model A) is treated generically in this paper; the actual formula, and any constitutional or statutory constraints on changing it, are a matter for FSM's national and state governments and outside this paper's knowledge.
• Model C's quarantine rule assumes political and legislative will to bind future windfalls in advance; FSM's own 2019 precedent shows the underlying fiscal judgement is achievable, but a standing rule is a different, and politically more demanding, commitment than a single discretionary decision.
• This paper does not model the monetary and financial-stability implications of a trust fund at 510% of GDP in a dollarised economy in any depth; that is a genuinely separate research question this paper flags without resolving.
• The long-running GRT reform recommendation (section 2.5) is described here at the level of public IMF reporting; the specific technical design questions involved in reforming FSM's gross-receipts tax are a matter for FSM's own tax-policy process and outside this paper's scope.

References

• International Monetary Fund. IMF Executive Board Concludes 2026 Article IV Consultation for the Federated States of Micronesia. Press Release 26/138, 27 April 2026.
• International Monetary Fund. IMF Staff Completes 2026 Article IV Mission to the Federated States of Micronesia. Press Release 26/028, 30 January 2026.
• International Monetary Fund. IMF Executive Board Concludes 2023 Article IV Consultation with Micronesia. Press Release 24/66, 4 March 2024.
• International Monetary Fund. IMF Staff Concludes Visit to the Federated States of Micronesia. January 2019 mission statement.
• U.S. Department of the Interior, Office of Insular Affairs. Second Five-Year Review of the Compact of Free Association, as Amended, for the Federated States of Micronesia.
• Parties to the Nauru Agreement (PNA). Vessel Day Scheme documentation.
• Methods: as elsewhere in this series — Hutton, E. (2017), The Revenue Administration Gap Analysis Program, IMF Technical Notes and Manuals 17/04. Method citations are from memory; verify before publication.