Beyond the First Ripple

1.1 The passive-recipient error

Policy memoranda often calculate as if households, firms and foreign actors simply receive a new rule. Yet a tax changes the return to avoidance; a benefit changes work, reporting or eligibility behaviour; a price cap changes supply and quality; a procurement preference changes bidding and relabelling; an investment concession changes the timing and accounting of projects. Rational-expectations economics may overstate foresight, but psychology does not restore passivity: people still respond, though through habit, present bias, imitation, fairness judgements, salience and rules of thumb.

1.2 The silo error

Budgets, sectors and ministries are administrative partitions, not causal boundaries. A hotel incentive can alter imports, bank collateral, land prices, farm labour, water demand, airlift, electricity investment, coastal exposure and the tax base. In a small state, many of these effects pass through the same utility, port, bank, technical cadre or foreign market.

1.3 The average-case error

An average expected gain can conceal an unacceptable correlated loss. A 5 percent chance of a mild administrative delay is not equivalent to a 5 percent chance of losing a visa waiver, a correspondent-banking route or a sole generating asset. Expected value alone also undervalues option preservation when harm is irreversible.

1.4 The budget-year error

Cash accounting records today’s subsidy and today’s revenue. It can miss contingent guarantees, future maintenance, later refinancing, off-budget obligations, ecological depletion and the political difficulty of removing a benefit. Discounting is essential, but a high discount rate can become a mathematical version of electoral myopia.

1.5 The implementation-as-footnote error

An intervention competes for procurement staff, engineers, inspectors, data systems, legal drafting, political attention and public trust. A nominally sound project can congest the very administration needed to make it sound. Implementation is therefore an endogenous part of the causal mechanism.

The result is policy-chain neglect: the institutional and cognitive tendency to stop at the direct output, or at most its fiscal total, before asking how actors adapt and how effects re-enter the system.

2.1 The economist and the psychologist

The economist asks how relative prices and constraints change. The psychologist asks which change people notice, how they interpret it and whether they postpone action. A household may underinvest in an efficient appliance because the upfront cost is vivid and future savings are abstract. A firm may resist a rule it considers procedurally unfair even when compliance is affordable. A policymaker may anchor to the announced project cost and escalate commitment after overruns.

The synthesis is neither perfectly rational agents nor irrational noise. It is heterogeneous adaptation: actors have incentives, limited information, unequal capabilities and psychologically shaped perceptions. Policy should change defaults, timing, finance, information and feedback - not merely the nominal price.

2.2 The game theorist and the electoral politician

Time-consistency theory warns that a promise useful today may become inconvenient tomorrow. Investors anticipate expropriation; voters anticipate broken commitments; officials anticipate policy reversal. Yet a rule rigid enough to be “credible” may be politically impossible after a hurricane, court decision or recession.

The practical answer is contingent credibility: clear goals, published formulas, independent measurement, narrow escape clauses and an announced route back to the rule. Grenada’s natural-disaster clause illustrates the principle. It was credible precisely because the exception and its trigger were specified in advance.

2.3 The manager and the political economist

Management asks who owns the process. Political economy asks who owns the bottleneck, who benefits from complexity and who can delay. A faster approvals target may improve service, or produce perfunctory scrutiny. A local-content quota may create supplier capability, or a market in relabelled imports. Every KPI creates a game.

The response is not metric avoidance. It is metric ecology: one outcome measure, one process measure, one distributional measure and one anti-gaming check, supported by spot audits and the authority to change the metric.

2.4 The financier and the development economist

Finance properly discounts future cash flows and distinguishes liquidity from solvency. Development economics asks whether today’s investment changes tomorrow’s productivity, skills and feasible choices. These lenses conflict most sharply when a policy creates a capability whose value is uncertain and long-lived.

RIPPLE-4 resolves the conflict by refusing a single scalar answer. Expected discounted value is reported alongside tail exposure, irreversibility, distribution and capability formation. Rights, basic safety and hard debt limits are constraints, not merely entries with small weights.

R - Reference the decision

Write the decision as an instrument, not an aspiration. “Increase food security” is an objective. “Require public institutions to procure 25 percent of specified foods through certified domestic aggregators, with a 30-day payment rule” is an analysable policy.

Record:

• the objective and success measure;
• the no-policy counterfactual;
• affected population and geography;
• start, review and end dates;
• non-negotiable constraints: rights, safety, treaty obligations, debt sustainability and service floors;
• the forecast vintage and evidence date.

I - Identify immediate incidence

Map who legally receives, pays, administers or is excluded. Then map who economically bears the result. A hotel tax may fall on owners, workers, suppliers or visitors depending on bargaining power and demand. A fuel subsidy benefits users in proportion to consumption, not necessarily need. A capital grant can increase supplier prices.

Use the double-entry discipline: every benefit requires a resource, liability, risk transfer or opportunity cost somewhere else.

P - Predict participant adaptation

For each material actor ask:

If this policy changes my costs, benefits, risks, information, status or sense of fairness, what becomes my best response - and what shortcut might I actually use?

Include households, firms, informal operators, civil servants, banks, unions, political parties, local communities, foreign investors, airlines, visa authorities, correspondent banks, donors and regulators. Identify who can comply formally while neutralising the policy in substance.

P - Propagate through coupled systems

Trace each important response through the seven RIPPLE ledgers:

1. Households and wellbeing: prices, income, time, safety, health and access.
2. Firms and production: costs, entry, exit, investment, innovation and supply chains.
3. Government: revenue, expenditure, debt, guarantees, staffing and legitimacy.
4. External account: imports, exports, foreign exchange, profits, remittances and mobility.
5. Finance: credit, collateral, liquidity, insurance and correlated balance sheets.
6. Institutions and politics: trust, compliance, rents, coalitions, litigation and credibility.
7. Natural and infrastructure assets: water, land, energy, ecosystems, maintenance and physical bottlenecks.

L - Locate leverage, lock-in and limits

Find the nodes through which several pathways pass. Look for:

• single-point infrastructure;
• scarce skills or thin supplier markets;
• common banks, insurers or public balance sheets;
• threshold effects and non-linear costs;
• contracts, construction or expectations that are expensive to reverse;
• recurring expenditure financed by volatile revenue;
• metrics vulnerable to gaming;
• dependence on foreign actors Saint Lucia cannot command.

E - Engineer evaluation and evolution

Convert the analysis into governance:

• a named accountable owner;
• leading indicators one causal order before the harm;
• data frequency and publisher;
• review dates;
• thresholds for continue, adjust, pause or stop;
• sunset and clawback provisions;
• authority and financing to execute the response.

A policy without a revision mechanism is a forecast pretending to be a commitment.

A map is not the deliverable

The test is whether a pathway changes one of the following: instrument, sequence, scale, eligibility, finance, contract, safeguard, metric, trigger, owner or timing. A beautiful diagram that changes none of these is analysis theatre.

6.1 Estimate conditional pathways

For a four-link chain:

P(E4) = P(E1) x P(E2 | E1) x P(E3 | E1,E2) x P(E4 | E1,E2,E3)

Do not multiply independent-looking numbers when effects share the same port, bank, ministry, visitor market or political trigger. Shared causes require joint scenarios or explicit correlation. When the foreign actor is strategic, estimate its best responses under different payoffs rather than treating its action as a random weather event.

6.2 Build estimates in this order

1. Reference class: What happened in comparable policies or states?
2. Preconditions: Which links must be true for the effect to occur?
3. Actor response: Who gains by making the effect more or less likely?
4. Scenario dependence: How does the probability change under optimistic, base and adverse conditions?
5. Independent elicitation: Where feasible, obtain estimates before group discussion to reduce anchoring and hierarchy effects.
6. Update rule: Name the evidence that will revise the estimate and when.

Probability and confidence are different. A pathway can have a low estimated probability but high confidence, or a wide probability interval because evidence is weak. Weak evidence should not force the weight toward zero. It should widen the interval, raise the value of information and favour staged action.

6.3 The expected system value screen

For pathway or scenario j, use:

ESV = sum[p(j) x M(j) x B(j) x H(j) x delta(tj)]

Where:

• p is the conditional probability;
• M is signed magnitude in a common or transparently scored unit;
• B is breadth - population and systems reached;
• H is persistence;
• delta(t) is the disclosed time treatment or discount factor.

Run the calculation under alternative distributional weights: equal social weight, pro-poor weight and, where material, a worst-case-regret view. Do not monetise away constitutional rights, minimum safety, severe ecological irreversibility or hard debt limits. Those remain constraints.

6.4 The critical-tail screen

Expected value can bury a low-probability catastrophe. Use the Critical Tail Index for triage:

CTI = sqrt(p upper) x Severity x Persistence x Irreversibility x Coupling x ControlDeficit

Score each non-probability factor from 1 to 5. p upper is the upper plausible probability, not the central estimate. The square root prevents an uncertain low probability from making catastrophic harm disappear. CTI is not a welfare formula; it decides which pathways require prevention, insurance, staging, redundancy or a stop rule.

6.5 Four decision postures

8.1 A formal intuition

Let the state vector be:

x = (H, F, G, E, B, I, N)

representing households, firms, government, external account, financial balance sheets, institutions and natural/infrastructure assets. A first-order policy impulse is d1. Actor adaptation and system transmission produce a scenario-specific interaction matrix J.

Up to four orders:

D4 = (I + J + J^2 + J^3) d1

If the effects of J dampen quickly, higher orders decay. If common nodes and reinforcing feedback increase the effective strength of J, ripples persist or amplify. The expression is a thinking aid, not a claim that Saint Lucia’s economy can be estimated as one stable linear matrix.

The four-order reverberation multiplier is:

M4 = ||D4|| / ||d1||

Comparing this multiplier across scenarios helps reveal when a small direct policy has large systemic consequences.

8.2 Small-State Recursion Density

A complementary diagnostic is Small-State Recursion Density: the weighted share of plausible four-order paths that return to a critical node such as the public balance sheet, utility, bank, port, tourism market or foreign-exchange pool.

RD4 = sum[path weight x re-entry indicator] / sum[path weight]

High recursion density warns that effects are not diversifying across independent systems; they are converging on the same capacity or balance sheet.

8.3 Feedback compression

The Small-State Feedback Compression Effect is the time dimension of the same structure. Openness, concentration and few intermediaries can make a third- or fourth-order effect arrive within one budget or electoral cycle. A foreign visa decision, airline route change, fuel-price shock or failure of a single asset need not wait for a long domestic chain.

8.4 Policy-chain neglect and modularity illusion

Two biases explain why coupling is missed:

• Policy-chain neglect: stopping after the visible output and perhaps the direct fiscal cost.
• Modularity illusion: assuming ministries, markets, firms and infrastructure are more independent than they are.

These are proposed concepts, not claims that the literature has never described related phenomena. They extend established work on systems dynamics, network contagion, policy resistance, common-pool resources, public-choice incentives and financial correlation into an operational small-state architecture.

Decision and context

Mauritius enacted its Export Processing Zones Act in December 1970. The regime offered duty-free imported inputs, tax concessions, repatriation rights and preferential access to European markets. It did not immediately dismantle protection for incumbent domestic firms. Sugar earnings, education, infrastructure, exchange-rate management, local commercial networks and foreign capital complemented the EPZ.

Four-order cascade

The pivotal second-order effect was not merely job creation. It was the mobilisation of labour, domestic capital and political support around a new export regime. The EPZ operated as a bounded reform laboratory. Incumbent protected firms did not all become immediate losers; this reduced opposition while the export coalition gained strength.

The Lewisian reading is exact: sugar rents mattered because of the use made of the surplus. The state and private sector combined rents with education, infrastructure and a trade regime. A tax holiday in isolation would not have produced the same chain.

RIPPLE-4 map

EPZ rules -> firm entry and women's formal employment -> skills, foreign exchange and exporter coalition -> higher wages and capabilities -> upgrading and diversification pressure

Hidden conditions: European preferences, stable institutions, trainable labour, local investors and complementary public investment.

Distribution: exporters and new workers gained; wage inequality burdened many women; later textile displacement required adjustment. Aggregate success did not remove the need to examine gendered incidence.

Decision-changing lesson for Saint Lucia: do not copy the instrument; assemble the transition system. An incentive should be appraised by the capabilities, domestic co-investment and constituency for upgrading that it creates. Require a route from preference or tax advantage to productivity before the advantage expires.

Sources: Subramanian and Roy (2001); World Bank (1989; 2013); Rogerson (1993); WTO (2015).

Decision and context

Beginning in 1974, Barbados combined duty exemption for inputs used in local solar-water-heater production with a 30 percent consumption tax on electric water heaters. In 1977, government required solar heaters in new public housing. From 1980, installation costs were tax deductible within a cap. Credit-union finance, loan guarantees, local manufacture, public leadership and product guarantees completed the package.

Four-order cascade

In 2002, reported fiscal incentives cost roughly US$550,000, about 0.06 percent of government revenue, while modelled annual consumer savings were more than twenty times that amount. The exact ratio is less important than the mechanism: a modest fiscal action mobilised household capital, local production and future energy savings.

The package affected both demand and supply. A rebate alone might have created imports and opportunistic installation. A mandate alone might have created resistance. Finance matched repayment to savings; guarantees made quality credible; public housing broadened access beyond income-tax-paying homeowners.

RIPPLE-4 map

Relative-price shift + public procurement -> adoption -> supplier capability and social proof -> fuel and capacity savings -> durable clean-energy constituency

Balancing loop: upfront cost and tenure limited the remaining market. Distributional warning: a tax deduction naturally favours tax-paying owners over poor tenants unless public procurement or targeted finance corrects it.

Decision-changing lesson for Saint Lucia: distributed energy is a market architecture, not a panel subsidy. Sequence hosting-capacity maps, resilient standards, tariff design, finance, installer certification, storage and low-income access. Anticipate the utility’s revenue and system-balancing response before scale creates conflict.

Sources: U.S. Department of Energy (2015); ECLAC (2019); CDKN (2012); IDB (2014).

Decision and context

In 1974 Jamaica imposed a bauxite production levy indexed to the realised price of primary aluminium and pursued majority public ownership of mining assets. The policy addressed a genuine incidence problem: vertically integrated multinational companies could influence transfer prices and retain resource rents abroad.

Four-order cascade

The result is policy-incidence migration. The law taxed Jamaican production. The economically relevant actor was the multinational’s global investment committee. A static model counted revenue on the existing base; a recursive model asked how the base would change.

This does not prove that the levy was a mistake or the sole cause of Jamaica’s macroeconomic crisis. Strikes, an alumina-plant explosion, oil shocks, wage and fiscal policy, and source diversification already underway also mattered. The World Bank’s more careful claim was that the levy helped make diversification away from Jamaica definite and difficult to reverse.

RIPPLE-4 map

Resource levy -> treasury windfall -> multinational portfolio reallocation -> output/export/reserve decline -> weaker fiscal and bargaining position

Strategic actor: the multinational’s portfolio, not the legally taxed mine. Lock-in: once new capacity is built elsewhere, a later tax reduction may not bring investment back. Possible redesigns: staged rates, capacity-utilisation incentives, stability agreements, regional producer coordination and explicit reinvestment conditions.

Decision-changing lesson for Saint Lucia: before counting a hotel, financial, telecommunications, land or citizenship levy, model how mobile actors can substitute across jurisdictions, contracts, transfer prices and time. Revenue projections require a behavioural tax base.

Sources: Government of Jamaica, Bauxite Production Levy Act; World Bank (1978); U.S. Department of State (1974); ODI (1981).

Decision and context

Tourism developed around a “one island, one resort” architecture. State ownership of land and the historic restriction of guesthouses on inhabited islands reinforced separation. The model provided privacy, controlled service quality and social distance between international visitors and inhabited communities.

Four-order cascade

The case is analytically powerful because the same feature - separation - had opposite effects. It generated privacy, quality control and a luxury premium. It also made it harder for women, local communities, farmers and small suppliers to participate. The correct question is not whether the resort model “worked” but which dimensions of success created the next vulnerability.

Ecological feedback completes the chain. Reefs, beaches and coastal systems produce the tourism premium; waste, reclamation and climate exposure can degrade that natural capital. The fiscal engine can consume its own asset base.

RIPPLE-4 map

Separated luxury product -> premium, FDI and fiscal rents -> national redistribution -> enclave imports and labour -> concentration, debt and ecological exposure

Success-induced constraint: expenditure expectations grew around a volatile revenue base. Distribution: the treasury and resort economy gained strongly, while domestic employment and supplier participation lagged gross output.

Decision-changing lesson for Saint Lucia: evaluate a resort incentive by retained domestic value per scarce unit - visitor, room, acre, cubic metre of water, kWh and dollar of tax expenditure - and stress it against airlift, source-market, disaster and ecological shocks.

Sources: World Bank (2015; 2022); Government of Maldives (2023).

Decision and context

After Hurricane Ivan, default and two restructurings, Grenada’s 2015 debt restructuring inserted natural-disaster clauses into its 2030 international bond and some official obligations. The bond allowed one or two semiannual payments to be deferred when modelled cyclone losses crossed specified thresholds, verified through the Caribbean Catastrophe Risk Insurance Facility.

Hurricane Beryl in July 2024 provided the first operational test.

Four-order cascade

Updated IMF reporting places Beryl’s damage and loss at about 17 percent of GDP. The clause did not remove this loss. It changed who supplied liquidity and when. This distinction is essential: liquidity relief is not solvency repair, and deferred amounts create a later repayment hump.

RIPPLE-4 map

Pre-agreed disaster clause -> verified trigger -> automatic payment pause -> response liquidity and lower default pressure -> later repayment hump + contractual demonstration

Design virtue: the second-order fiscal response was specified while bargaining power remained intact. Control: an external parametric verifier reduced discretionary conflict. Tail guard: debt limits and a post-event repayment plan remain necessary.

Decision-changing lesson for Saint Lucia: predictable disasters belong in financing contracts, project design and fiscal rules before they happen. Pair insurance, contingent credit, cash reserves, resilient infrastructure and debt clauses; measure the combined timing of payouts and later obligations.

Sources: IMF (2017; 2022; 2025a); Federal Reserve Board (2021); World Bank and CCRIF documentation.

Decision and context

Vanuatu expanded modern investor-citizenship schemes from 2015. Passport sales generated exceptionally large public revenue. The product’s value depended partly on visa-free mobility granted by other jurisdictions. High application volume and weak screening raised security and migration concerns outside Vanuatu.

Four-order cascade

This is a recursive common-pool problem. Government sells citizenship because mobility is valuable. Weak controls and high volume change the foreign gatekeeper’s risk calculation. The gatekeeper withdraws mobility. The programme then devalues the asset on which its own demand depends, while imposing costs on citizens who never participated.

Dominica provides the useful positive-side comparison. Its citizenship revenue helped finance post-disaster resilient housing and public investment. Yet IMF analysis likewise warns that volatile receipts should fund capital, debt reduction and saving rather than politically sticky recurring expenditure. Productive use of the windfall does not eliminate the external gatekeeper risk.

RIPPLE-4 map

Citizenship sales -> fiscal windfall and programme dependence -> screening/reputational externality -> foreign visa action -> universal mobility cost + programme devaluation

This paper calls the universal spillover the Mobility Collateral Effect: risks generated by monetising citizenship are socialised across all passport holders, while the immediate fiscal benefit is controlled by the programme and government.

Decision-changing lesson for Saint Lucia: the UK, EU, banks and security partners are actors inside the programme’s theory of change. A prudent allocation rule cannot substitute for due diligence, data sharing, regional standards and credible volume control; both fiscal and reputational safeguards are required.

Sources: IMF (2017; 2023; 2025b); Council of the European Union (2024); European Commission (2024); Australian Department of Foreign Affairs and Trade (2022).

What generalises - and what does not

No country supplies a plug-and-play policy. Mauritius had trade preferences, networks and institutions that another EPZ may lack. Barbados combined several instruments over decades. Jamaica faced a vertically integrated global industry. The Maldives’ geography enabled a product Saint Lucia cannot reproduce. Grenada’s clause works as one layer, not a substitute for fiscal space. Vanuatu’s experience demonstrates a mechanism, not an inevitable future for every citizenship programme.

What generalises is the method: identify complementary assets, strategic responses, common nodes, feedback, lock-in and distribution; then redesign the decision around them.

18.1 Forecast-vintage risk is itself a policy lesson

The 2025 growth estimate demonstrates why a Cabinet paper must date its baseline. The IMF’s January 2026 report projected 1.7 percent growth for 2025; the World Bank’s April 2026 update estimated 1.3 percent; the later official review reported a preliminary contraction of 0.6 percent. These are different vintages, not interchangeable truths.

RIPPLE-4 therefore requires:

• a publication date beside every macro assumption;
• a range rather than a single point for trigger design;
• automatic re-estimation when an outturn replaces a forecast;
• an explanation of whether the decision survives the revision.

19.1 Bananas - preference erosion, scale and concealed distribution

In the 1980s, agriculture - mainly bananas - accounted for roughly one-fifth of GDP, half of merchandise exports and close to one-third of employment. Changes to the European regime reduced returns; the 1997 WTO ruling deepened preference erosion. By 1999-2002, bananas still represented 39.5 percent of merchandise exports and involved workers equal to 10.8 percent of the working-age population. Tourism growth helped replace aggregate export earnings. It did not automatically replace the rural production system.

Preference erosion -> lower unit earnings -> farm exit and lost packing/shipping scale -> labour and capital shift -> greater tourism concentration + weaker food capability

The third- and fourth-order effects matter. When volume falls, unit shipping, marketing and quality-control costs can rise. That induces more exit: a reinforcing cost spiral. Underused land, lost cooperative capacity and rural inequality can persist even while national GDP recovers. Later local-food initiatives then begin from a thinner supplier base.

Lesson: evaluate adjustment at two levels - the national aggregate and the incidence map by district, gender, income, occupation and generation. “The economy diversified” can be true while “the adjustment was resilient and inclusive” is false.

19.2 Tourism - growth, retained value and correlated exposure

Tourism expansion produces a familiar direct chain: construction, rooms, visitor spending, jobs and tax. Its higher orders depend on how much value remains and what scarce systems it binds.

More rooms and airlift -> spending and jobs -> imports, water, energy, land and skill demand -> supplier gains or leakage/congestion -> national exposure to source markets, airlines and coast

A World Bank benchmark for six OECS economies found that hotels, cruise operators and marinas sourced only around 32 percent of food demand locally at the time studied. The estimate is dated and not a current Saint Lucia ratio, but its diagnosed barriers remain instructive: volume, consistency, standards, logistics and aggregation. A procurement target without those capabilities may create relabelling, shortages or higher prices rather than farm upgrading.

The correct headline is therefore not arrivals alone. It is retained domestic value per scarce unit: per visitor, room, acre, cubic metre of water, kWh, tonne of imported input and dollar of tax expenditure.

19.3 Citizenship by Investment - a realised fourth-order mobility effect

The programme produces immediate fiscal and real-estate receipts. It can finance debt reduction, resilience or productive investment. It can also create project dependence, processing pressure and reputational externalities.

Applications and fees -> fiscal space/projects -> volume, screening and reputation signals -> foreign gatekeeper review -> mobility cost + programme value change -> future fiscal adjustment

On 5 March 2026, the United Kingdom imposed visit-visa and direct-airside-transit-visa requirements on Saint Lucian nationals, with a transition period to 16 April. The UK cited asylum claims and the scale and misuse risks associated with citizenship sales among its stated reasons, including the rise to 5,642 applications. That does not establish CIP as the sole cause. It does establish that programme design became one input into an external sovereign decision affecting every Saint Lucian traveller.

Revised EU rules effective in late 2025 also permit visa-free suspension where investor citizenship is granted without a genuine link. As at the date of this paper, this is a legal exposure, not an EU suspension of Saint Lucia.

This is the Mobility Collateral Effect in realised form: a programme controlled by a narrow set of actors alters a national asset used by all citizens. If lost mobility reduces future demand, the programme can partly consume the value it sells.

19.4 Renewable energy - option value, grid response and distribution

Saint Lucia’s renewable-electricity ambition is strategically coherent: imported fuel, a concentrated generation site and climate exposure create strong reasons to diversify. The target should be stated with its source - 50 percent by 2030 in the 2023 National Energy Policy, and 40 percent by 2030 with 46 percent by 2035 in the 2025 NDC. The pathway is not “install renewable capacity, then save fuel.”

Renewable investment -> imported capex + installation -> lower fuel use and changing load curve -> utility revenue/tariff and grid-balancing response -> resilience or backlash/curtailment -> new skills, assets and liabilities

Geothermal exploration may reveal valuable baseload potential, but potential is not a committed 30 MW plant. Treat exploration as a real option with stage gates: geological evidence, drilling result, environmental safeguards, finance, tariff impact and exit criteria. Distributed solar likewise requires feeder hosting maps, storage, time-sensitive pricing, hurricane-resistant standards, installer capability, low-income finance and a utility remuneration model. Without these, adoption can shift fixed network costs onto renters and households least able to install.

Lewis’s agrarian-industrial warning applies by analogy: the enabling system and the new industry must advance together. Generation without grid capability can become its own bottleneck.

20.1 A new hotel tax holiday or resort incentive

First-order case offered: foreign investment, construction, rooms, jobs and future visitors.

Metrics: retained visitor dollar; local food and professional-service share; capex import content; median Saint Lucian wage; tax expenditure per sustained job; water and kWh per guest-night; insurance coverage; airlift load factor.

Adapt trigger: redesign or suspend benefits if local-value targets miss two successive reporting periods, a water-stress threshold is crossed, insurance lapses or the fiscal NPV becomes negative under the agreed arrivals shock.

20.2 Allocating a CIP windfall

First-order case offered: cash without conventional taxation and below-market or interest-free financing.

The policy must survive five mandatory scenarios: zero receipts for three years; a 30 percent application fall; EU visa suspension; a project delay that doubles cost; and a hurricane immediately after the revenue shock.

Use a transparent three-bucket rule:

1. Debt and liabilities: reduce expensive debt and disclose CIP-linked obligations.
2. Liquidity and disaster reserve: hold genuinely liquid, separately governed assets.
3. Productivity and resilience capital: finance only appraised projects that pass procurement, benefit-realisation and maintenance gates.

The structural budget should use only a conservative rolling-average draw. Recurring wages, open-ended transfers and routine operating costs should not depend on the current application surge.

Metrics: recurrent spending financed by CIP; sovereign liquid assets; debt service to current revenue; project completion and realised benefit; applications, grants, denials and processing time; escrow and land-sale transparency; external mobility warnings.

Stop/adapt trigger: an external formal warning, material due-diligence breach, unexplained escrow discrepancy or reserve falling below the stated shock requirement automatically reduces programme-dependent commitments and triggers independent review.

20.3 From general fuel subsidy to targeted relief and distributed energy

General price support provides immediate relief. Its second order can be higher consumption and a benefit distribution tilted toward larger users. Its third order is less fiscal room for cash support, efficiency, grid investment and storage; its fourth order is fossil lock-in and abrupt adjustment when oil or fiscal conditions change.

The alternative also has a cascade. Solar and storage initially require imported capital and skilled installation. They alter utility revenue and load. If tariffs and hosting limits lag, cost shifting, curtailment and backlash follow.

Design: protect low-income households through targeted cash or lifeline LPG; publish a rule-based subsidy glide path; pilot feeder-level PV and battery storage; publish hosting maps; introduce time-sensitive tariffs only with bill-distribution analysis; create a utility transition mechanism; retain geothermal as a stage-gated option.

Metrics: subsidy by income group; fuel-import bill; subsidy as GDP and revenue; peak demand; reliability; hosting capacity and curtailment; arrears; bill distribution; local technical jobs; avoided diesel.

Pause trigger: do not scale distributed generation on a feeder when hosting or protection studies are incomplete, unmanaged curtailment exceeds the approved level or the tariff transition creates unmitigated hardship in the bottom income groups.

20.4 The 2026 water tariff as a credibility compact

The National Utilities Regulatory Commission’s final determination took effect in June 2026. The lowest domestic base remains protected, while other domestic, commercial, government, hotel and boat tiers rise in phases. Part of the increment is earmarked for the John Compton Dam pipeline, with quarterly reporting and performance standards.

Higher tariff -> revenue -> repair/capital delivery -> lower loss and better reliability -> willingness to pay -> utility and fiscal sustainability

The loop can run backward:

Higher bill + delayed service -> distrust/arrears/illegal connections -> revenue loss -> weaker maintenance -> worse service

With reported non-revenue water as high as 58 percent, the lag between today’s bill and tomorrow’s service is the decisive political-economy node.

Design: publish a quarterly account and project dashboard; independently verify non-revenue-water, outage and capital milestones; protect vulnerable households through targeted affordability support; pre-authorise temporary bill credits or corrective plans when controllable milestones are missed.

Metrics: non-revenue water; interruption by district; repair time; water quality; arrears and disconnections; tariff burden by income; capex execution; energy per cubic metre; hotel water per guest-night.

20.5 A local-food procurement compact

The attractive first-order story is a guaranteed market for farmers. The real mechanism requires demand aggregation, product standards, crop planning, working capital, cold chain, timely payment and protection against false local content.

Procurement commitment -> predictable orders -> finance and farm investment -> aggregation/quality capability -> hotel and institutional substitution -> resilient food system

The negative loop is equally plausible:

Quota without capability -> shortages/price pressure/relabelling -> buyer resistance and late payment -> farmer cash stress -> policy reversal

Design: begin with a narrow basket where local production has comparative feasibility; certify origin; publish forward demand; use an aggregator with transparent margins; guarantee payment within 30 days; finance working capital; measure on-time-in-full delivery and buyer quality acceptance. Expand only after two production cycles meet the thresholds.

Metrics: verified local share by product and value; farmgate share of final price; on-time-in-full delivery; rejection and spoilage; payment days; participating small farms by district and gender; import displacement net of new imported inputs.

21.1 A two-page recursive impact statement for Cabinet

Every material Cabinet submission should include:

1. precise instrument, counterfactual and evidence vintage;
2. one four-order RIPPLE Map;
3. incidence across the seven ledgers;
4. the three highest-value pathways and three highest critical tails;
5. actor best responses, including one foreign gatekeeper where relevant;
6. distribution by income, gender, district, firm size and generation where material;
7. owner, indicators, review date, correction and stop rule;
8. a short record of how this analysis changed the proposal.

The aim is not a longer Cabinet paper. It is a harder, more decision-relevant one.

21.2 Institutional roles

21.3 A 30/90/365-day adoption route

Within 30 days

• adopt the one-page effect card for three pending decisions;
• name a central methodology owner;
• define probability and evidence labels;
• require forecast date/status on all macro figures.

Within 90 days

• run cross-ministry mapping sessions on one fiscal, one infrastructure and one regulatory choice;
• create a register of common nodes: banks, utilities, port, airport, scarce skills, foreign gatekeepers and public guarantees;
• publish the first trigger dashboard;
• train analysts to distinguish facts, causal inference and scenarios.

Within 365 days

• integrate the recursive impact statement into Cabinet and public-investment templates;
• back-test ten past decisions to create a Saint Lucian reference class;
• link review triggers to budget releases, concessions and contracts;
• require post-implementation maps to be updated with observed actor responses;
• publish an annual “Policy Ripples” review of which forecasts were right, wrong and revised.

21.4 AI as analyst, not accountable authority

An artificial-intelligence system can expand pathways, retrieve reference classes, surface contradictions and run scenarios. It is especially useful as a structured red team. It can also fabricate causal certainty, duplicate one benefit under several labels and bury value judgements inside weights.

Any AI-assisted RIPPLE-4 analysis should therefore disclose sources, data dates, confidence, assumptions and scenario ownership. Independent human officials remain accountable for rights, distributional weights, risk tolerance, evidence acceptance and the final decision. The machine may propose a stop rule; only an authorised institution can bind itself to act on it.

The 30-minute protocol

Ten dangerous blind spots

Feedback-loop typology

Systemic adjustment or critical tail?

Ask four questions:

1. Does an accounting identity, physical constraint or legal rule make the effect difficult to avoid?
2. Does an affected actor have both a clear incentive and the capability to respond?
3. Has the adjustment appeared in a relevant reference class?
4. Does it occur under optimistic, central and adverse scenarios?

If at least three answers are yes, put the effect in the baseline as a systemic adjustment, not a risk footnote.

A critical tail has a low central estimate but a meaningful upper probability and combines severe harm with irreversibility, concentrated exposure, correlated losses, weak early warning or low government control. Prevent, cap, insure, stage or preserve an exit option. A severe systemic adjustment and a critical tail can coexist.

Five questions to ask aloud

1. What will people do because this policy exists?
2. Where will the cost migrate?
3. Through which shared institution or bottleneck will the effect return?
4. Which external actor can rewrite the payoff?
5. What observable evidence would make us change course?