What's new at the desk.

Rule-taker by choice

The thirty-first country decision case examines Turkey's 1996 customs union with the European Union - adopted as Decision No 1/95 on 22 December 1995, in force 31 December 1995 (Article 65), implemented from 1 January 1996: the common external tariff adopted, bilateral rules of origin scrapped, EU regulation mirrored, agriculture and services excluded. Goods exports rose from 23.0 billion dollars in 1996 (WITS) to a record 262 billion in 2024; bilateral EU-Turkey trade reached a record 217.6 billion euros in 2025 (European Commission). The World Bank's 2014 evaluation credits the tariff anchor, origin-free trade and the regulatory mirror - the export series itself is reported as correlation, not causal identification - while the asymmetry's price is tracked through the EU-Mexico trade-deflection example and the modernization proposal unadopted since December 2016. Independent desk research, not peer reviewed and not investment advice.

Trading independence for self-government

The thirtieth country decision case examines the 2005 Aceh settlement - talks agreed days before the tsunami, five Helsinki rounds trading independence for self-government: 840 weapons surrendered, 24,000 troops withdrawn, 1,424 prisoners freed within five months, and former rebels contesting every election since. The security half of the bargain has held for twenty years; the promised human-rights court and full truth commission largely have not arrived, and contested implementation figures are carried as attributed assessments. Independent desk research, not peer reviewed and not investment advice.

Six years at full speed

The twenty-ninth country decision case examines New Zealand's 1984-1991 reform program - the dollar floated on 4 March 1985, farm subsidies that had reached 35 percent of pastoral output by 1983 abolished within twelve months, government trading departments corporatized under the State-Owned Enterprises Act 1986, and the world's first legislated central-bank inflation target through the Reserve Bank of New Zealand Act 1989 and the March 1990 Policy Targets Agreement. The record is double-sided: inflation fell from nearly 19 percent in 1987 to about 2 percent by 1991, and the framework legislated from 1989 still binds policy in 2026 through successor statutes - the 2021 Act replaced the 1989 Act and the 1994 fiscal-responsibility principles now sit in the Public Finance Act - while unemployment rose from 3.6 percent in 1986 to about 11 percent by 1992, with job-loss and poverty tallies carried as attributed, contested estimates. Independent desk research, not peer reviewed and not investment advice.

The asymmetry was the design

The twenty-eighth country decision case examines Sri Lanka's first bilateral free trade agreement - signed with India on 28 December 1998 and in force from 1 March 2000 - and the deliberately asymmetric tariff architecture at its core: India opened 1,351 tariff lines at zero duty immediately and finished its phase-out in three years, while Sri Lanka kept 1,180 lines on its negative list and phased cuts over eight. The case follows twenty-five years of measured use - exports to India from 49 million dollars in 1999 to over one billion in calendar 2025, with roughly sixty percent using the agreement's preferences against 9.4 percent on the Indian side - the vanaspati surge and quota settlement that tested the treaty, and a deepening twice negotiated and never signed: CEPA initialled in July 2008, and ETCA at a standstill as of September 2025, the last date-bounded status located. Independent desk research, not peer reviewed and not investment advice.

The room that Ceylon helped to build

The twenty-seventh country decision case examines the April 1955 Bandung Conference: five Colombo Powers - Burma, Ceylon, India, Indonesia and Pakistan - convened twenty-nine Asian and African governments representing 1.5 billion people, and closed the week with the unanimously adopted ten-point Dasasila Bandung. The case follows Ceylon's convening role, Sir John Kotelawala's undated Soviet-colonialism intervention (quoted only from his 1956 participant memoir and flagged as such), and the lineage to the Non-Aligned Movement - a founding seat at Belgrade in 1961 and the chairmanship exercised from Colombo in 1976. Independent desk research, not peer reviewed and not investment advice.

The package deal

The twenty-sixth country decision case examines Israel's July 1985 emergency stabilization: a budget cut of 7.5 percent of GDP, a devaluation of about 20 percent followed by a fixed rate, a wage agreement signed by the Histadrut and the employers after a one-day general strike, and a 1.5-billion-dollar American grant - recorded annual inflation fell from 449.7 percent in 1984 to 19.7 in 1986 without a recession. Transfer limits for Sri Lanka are stated before conditional lessons, with the annualized 500-percent launch pace kept distinct from the recorded annual series. Independent desk research, not peer reviewed and not investment advice.

The target was the test

The twenty-fifth country decision case examines South Korea's export discipline, 1961-1979: firm-by-firm export targets, a monthly president-chaired review from January 1965, an exchange rate realigned for exporters in 1965, and the heavy-industry bet that followed - exports rose 368-fold across the era. Transfer limits for Sri Lanka are stated before conditional lessons, with the SCM Agreement Annex I(i) boundary and a legal-review caveat named. Independent desk research, not peer reviewed and not investment advice.

Consent, not victory

The twenty-fourth country decision case examines Northern Ireland's Good Friday Agreement and the long maintenance of peace, 1998-2024: a consent principle that parked the final-status question, verified decommissioning witnessed by two clergymen, 428 licensed prisoner releases, policing rebuilt to 30 percent Catholic officers, and institutions that stopped four times and were rebuilt each time. Transfer limits for Sri Lanka are stated before conditional lessons. Independent desk research, not peer reviewed and not investment advice.

Too big to save

The twenty-third country decision case examines Iceland's banking collapse, capital controls and the long way back, 2008-2017: three banks near ten times GDP failing in one October week, the government's choice not to bail them out, eight and a half years of capital controls, Western Europe's first IMF program since 1976, and a recovery that reached full employment by 2016. The five-page reader and PDF keep measured outcomes beside the institutional caveats; Sri Lankan lessons stay conditional. Independent desk research, not peer reviewed and not investment advice.

The peace that came with a tariff line

The twenty-second country decision case examines Jordan's Qualifying Industrial Zones, 1994-2025: duty-free access to the United States conditioned on Israeli content, the South Asian migrant workforce that sewed the exports, the 2006 labor scandal and the monitoring rebuild that followed, and the 2025 tariff that landed despite a free trade agreement. The five-page reader and PDF keep measured outcomes beside labor and concentration limits; Sri Lankan lessons stay conditional. Independent desk research, not peer reviewed and not investment advice.

A quieter first screen

The home page now opens calmer: secondary navigation sits under one More menu, the language bar opens on demand (Sinhala and Tamil previews keep a short note), the key-figures strip is static until you press Play, one primary action leads, and the six audience views stay visible as quiet outlined links. No content, board or feature was removed; the terminal is unchanged.

From coffee to catheters

The nineteenth country decision case examines Costa Rica's investment-promotion machinery and export transition, 1982-2024. The eight-page reader and PDF keep measured export outcomes separate from claims about causes, education mandates separate from actual spending, and Sri Lankan lessons conditional. Independent desk research, not peer reviewed or investment advice.

Two slim, pausable tickers

The six dated figures now scroll in one thin row directly beneath the Morning Desk Note. Further down, a separate news strip links Sri Lankan and world headlines to their original publishers, with publication times. Both strips have pause controls and a static reduced-motion alternative. News is checked during the daily refresh and hidden after 36 hours if it has not been refreshed. Existing boards, forecasts and alert-send restrictions are unchanged; the wider numerical audit remains open.

Six task-specific views, with the full desk preserved

New pages organize the same dated desk for everyday readers, corporate users, economists and analysts, investors and brokers, policy and research users, and diaspora/remittance watchers. A full-data page keeps every existing board and model available. A separate major-shock alert page publishes draft triggers and consent rules; signup and sends remain pre-launch while the dedicated email service, double opt-in and unsubscribe path are verified. Frozen forecasts do not trigger alerts, and the wider numerical audit remains open.

Clearer port-board column labels

The mobile Port of Colombo table gives the planned-moves column more room so its heading stays separate from vessel length. Desktop layout and source data are unchanged. This is a narrow readability fix, not a numerical audit clearance.

A stock watchlist and like-period bank comparison

My feed now has a watchlist for the existing 28 covered companies, with session-dated quote snapshots and selected macro releases linked to primary sources. Guest choices save in this browser; signed-in choices use a separate account watchlist. Save errors remain visible. Bank comparisons use standalone-bank H1 2026 profit and reported ratios for COMB, HNB, Sampath and Nations Trust, with the same period and perimeter, CSV export and original statement pages. Definitions and one-off effects remain explicit. No trading, holdings or alert delivery is added.

The dated stock-history foundation is prepared, but the live archive is empty and collection is disabled until Colombo-source storage and redistribution rights are established. No historical equity chart or backfill is claimed. The wider numerical/history audit and the earlier company-landscape/BOI work remain open.

Derived reports and freshness checks reconciled

Desk reports now show the existing frozen 4.6% GDP nowcast rather than a stale 5.0% copy, and the no-script rupee gauge was verified against its published inputs at 47/100. The freshness register now watches the dated event board and recognises the procurement check stamp. Food 13 Oct prices and source dates are updated. Ask the Desk now dates the gauge inputs as September rather than repeating an old August label. The official-outlook comparison now uses the 13 Oct IMF WEO (GDP 4.0%/3.7%, annual-average inflation 5.3%/5.4% for 2026/2027), with the older May numerical snapshot retained. Forecast ranges and model formulas are unchanged; The regional growth/per-capita comparison now uses the same October WEO vintage; the mechanical eight-peer set is unchanged, with the April table retained. The official-outlook page now matches the warm ivory site background; the desktop header separates More from search, and the mobile food table has a visible sideways-scroll hint and keyboard focus. The broader primary-source and history audit remains open.

Mauritius: an export window is not a permanent advantage

The thirteenth country decision case examines the export-processing bargain and adjustment, 1970-2005. Its eight-page reader and PDF distinguish input relief, market access, worker outcomes and export measures. Employment fell 25.6% between December 1999 and December 2004 in the cited EPZ series; that is not a causal estimate of quota expiry. Sri Lankan lessons remain conditional. Earlier cases remain available. Review draft, not peer-reviewed research or investment advice.

Foreign Investor Dashboard: sector-specific institutional screens

The dashboard now lets readers choose twelve sectors. Each selection changes the disclosed governance weights, score and driver while retaining the country baseline for comparison. These are analyst assumptions drawn from the Sri Lanka transmission dossiers, not calibrated investment-loss risks or proof that a sector is safe. Market and project gates remain unscored. Country data, evidence dates and peer comparisons are unchanged.

Investor terminal: country indicators and peer comparison

A warm-ivory terminal now compares Sri Lanka with Singapore, India and Vietnam on five dated governance measures and FDI inflows. The visible weighted score is an institutional screen, not capital safety. Risk bands, uncertainty intervals, missing project gates and exact country data are disclosed. Investors is in the main navigation.

The same reserve inputs across the gauge and its dossier

The rupee-pressure dossier now agrees with the visible gauge: end-September reserves of $6.851bn, down 1.1% from revised end-August $6.929bn. Its fallback score is 47, WATCH. The methodology now describes the implemented FX move against the stated end-July baseline, rather than a rolling 30-day move. No formula or forecast changed. The exports board's current-account row is correctly labelled August, not July. The wider numerical audit remains open.

Camp David: two frameworks, unequal outcomes

The ninth country decision case follows the 1978 frameworks, the 1979 Egypt-Israel treaty and Sinai monitoring in 1982. Its seven-page reader and PDF keep bilateral implementation separate from the unresolved Palestinian track, domestic approval and absent-party consent. Sri Lankan lessons remain conditional. The eight earlier cases remain available. Review draft, not peer-reviewed research or investment advice.

A personal first view, with the full desk still open

Signing in now opens My feed. Choose macro topics and optional sector shortcuts; preferences save to your account and stay separate from daily-email settings. Guests can try the controls without signing in. Every existing board, model and country case remains available. A separate news page brings recent Newswire publisher-feed headlines with dates, attribution, links and our own desk-context notes, not republished articles. Automated selection favours current Colombo-day headlines and macro-topic matches. Publisher data refreshes through the site's own API on request, with an hourly cache. Ada Derana is not included while its direct source cannot be verified. Save errors are visible and keep the choices in the current preview; live cross-device persistence remains to be confirmed with a real account. No news item changes a model automatically.

Try My feed · Read recent headlines

GDP history: release vintages made explicit

The stored GDP history combines release vintages rather than a single current revision round. Q2 2025 growth of 4.9% matches the original DCS release; the current Q2 2026 release revises that comparison to 5.0%. The library and nowcast depth note now state the limitation. The current Q2 2026 headline remains 4.2%. No historical values, frozen forecasts or model sensitivities changed. A revision-harmonised back-series remains part of the open audit.

Read the GDP history note

Policy rates: separate regimes and a dated current snapshot

The library now separates the pre-November 2024 SLFR policy-rate history from the Overnight Policy Rate series. Standing facilities still exist, but ceased to be policy rates on 27 November 2024. The latest snapshot is explicitly the 30 September review: OPR 8.75%, SDFR 8.25%, SLFR 9.25%. Existing monthly history still ends in July; no missing monthly points are inferred. The OPR observation count is corrected. Forecast ranges and model sensitivities are unchanged; the wider number audit remains open.

Read the corrected policy history

Services PMI: one measure, one window

The services history now shows the Business Activity Index from January 2023 to August 2026: 44 monthly observations checked against CBSL MEI Table 10. The older composite PMI is withheld rather than joined to a different measure. Business activity became the services headline in January 2024; earlier business-activity subindex observations remain comparable. Its latest percentile is now 77th within this shorter window. Shared historical chart high/low labels now use the actual observations, not padded plotting bounds. Latest inputs, forecast vintages and model sensitivities are unchanged; the wider number audit remains open.

Read the corrected history library

Credit history: corrected measure and reference months

The historical line previously called private credit was domestic-banking-unit total loans and advances. June is 24.7% and July is 24.2%, verified from CBSL monthly indicators. Twelve consecutive reference months, August 2025 to July 2026, have now been re-derived from the primary tables, including the missing March reading of 23.8%. Older mislabelled and month-shifted points remain withheld while their reference periods are checked. This is separate from the current banking-system private-credit print of 24.5% for August. The monthly 364-day T-bill series now includes August at 9.89%, the last auction of that month. No forecast or model sensitivity changed; the full-number audit remains open.

Read the corrected ledger

Bank size, without misleading market-share ratios

The six-bank board now uses primary standalone H1 financial statements at 30 June. The old rounded Q1 percentages and summed concentration estimates are removed. Deposits and gross customer loans are shown separately, with issuer sources and geography/category limits. CBSL Q2 industry and LCB totals are context, not silent ratio denominators. The full-number audit is still open.

Read the dated bank board

Canada: a number is not yet credibility

The eighth country decision case examines Canada’s inflation targets, 1991-2001. Its seven-page reader and PDF separate announced targets, policy tools and measured inflation from claims about causation, credibility and adjustment costs. Lessons for Sri Lanka remain conditional. The seven earlier cases remain available. Review draft, not peer-reviewed research or investment advice.

Food at 8 October, labour at Q2

The terminal food basket now reflects the CBSL 8 October daily print, including Dambulla pumpkin up 15.4% and snake gourd down 3.8% against 7 October. Labour unemployment and participation are now 4.0% and 50.0% for Q2 2026, explicitly attributed to CBSL's 2 October weekly report while the DCS quarterly index still lists Q1. Historical annual rows keep their dates. These are targeted corrections, not full-terminal numerical clearance.

Policy and legal evidence, with date limits visible

The policy and legal register adds eight certified Acts with topic and evidence filters, search, source links and CSV downloads. A separate Inland Revenue matrix keeps 23 commencement/application rows apart. Missing Gazette orders and exceptions remain explicit; a certification date is not presented as proof that every provision is operative. This is a bounded source register, not legal advice or a complete compliance audit.

Vietnam and the US: normalisation, trade and industrial upgrading

The seventh country decision case follows Vietnam-US normalisation and trade access. Its seven-page reader and PDF separate documented trade outcomes from mechanisms and conditional lessons for Sri Lanka. The six earlier cases remain available. This is a review draft, not peer-reviewed research or an official forecast.

Fiscal flows and gross debt, kept separate

The fiscal and debt ledger adds January-August revenue/spending, annual-estimate execution, PDMO's end-June gross debt composition in two currencies, domestic instruments and Q2 service. Treasury/PDMO documents and CSV downloads sit beside definition notes. Different CBSL debt coverage and unresolved source arithmetic are explicitly marked, not forced to match. This is a bounded fiscal snapshot, not a full debt or legal audit.

Official outlooks, with the vintages visible

A new official outlooks page compares IMF, ADB and World Bank annual growth and average-inflation views for 2026 and 2027. Sources, publication dates and definitions stay alongside each number. The IMF seventh-review staff-level agreement is shown separately from conditional financing access. The frozen Forward view and quarterly nowcast are unchanged; no consensus or forecast-skill claim is made.

Power, from daily mix to plant readings

The power explorer adds a verified 30-day dispatch window, a daily source selector, history chart and table, searchable plant readings and CSV downloads. The 6 Oct total and mix reconcile with PUCSL and CBSL. Estimated sources and missing storage are explicit. Costs remain withheld because the publisher's public cost display returned zero without a verified reconciliation. This is generation coverage, not full power-sector parity.

Rates and prices, with the periods kept apart

A new rates & prices page separates CCPI from NCPI and headline from core inflation. It includes 13 printed history months per index, a policy/funding/lending ladder, 24 published bank AWPR readings with search and sorting, five liquidity days and selected secondary bill quotes. CSV downloads and the structured snapshot carry the dated source readings. The GDP nowcast explanation now agrees with its +4.6% headline; the frozen Forward view is marked separately from the newer CBSL inflation projection. This is a first source-grounded release, not a full numerical clearance of the terminal.

Six country decision cases

The research library now contains decision cases on Singapore, India, the Philippines and Malaysia, plus the US-China diplomatic opening and a Sri Lanka-Singapore growth comparison from 1950. Cases separate documented outcomes from mechanism hypotheses and conditional lessons for Sri Lanka. All remain review drafts, not peer-reviewed research or official forecasts. The landing page now has a direct Research button. Social previews and captions are no longer on the site.

Jump between desk boards, and a bigger Generate report control

A "Jump to" bar above the live desk and the sector desks links straight to the rupee, port, event log, desk read, commodity, T-bill, hydro, power, trade, tourism, food and procurement boards, and it stays in view as you scroll. The Generate report buttons (Weekly, Monthly, Q3 2026, Yearly) are now large, labelled and easy to tap on a phone.

Today and the outlook, at the top of the terminal

A new ledger at the top of the terminal lists the latest dated print for ten headline indicators (policy rate, inflation, USD/LKR, reserves, private credit, exports, trade balance, remittances, tourist arrivals, GDP growth) in three groups, each row with the change on the prior print and a labelled history line. Where a Forward view range exists (five indicators, 15 Sep 2026 vintage), it sits on the same row with a marker showing whether the latest reading is inside, above or below it. Ranges are conditional and kept apart from measured data; they are not point forecasts. Rows animate in and respect reduced-motion settings.

Neighbour exposure: an interactive map of India, China and peers

A new Regional board shows how much of Sri Lanka's tourism and trade runs through India and China, and compares Sri Lanka's real GDP growth with its neighbours and with a rule-based peer group of Asian economies at a similar income level. Select a country on the map or in the table for its growth, the gap to Sri Lanka and any verified links. A "what could this mean for Sri Lanka?" explorer lists the channels (tourism, exports, imports, currencies, rates) through which a change in India, China, the United States or the United Kingdom could reach Sri Lanka, with measured exposure kept apart from mechanism reasoning and no impact estimates. Every figure carries its publisher and date, the peer rule is published on the board, and missing data is shown as missing. It is descriptive only: it does not claim that neighbour activity causes Sri Lankan outcomes.

Get the daily brief with Google

The daily-brief account page now offers Google sign-in alongside email and password. Only basic profile and email details are requested. The terminal stays open without an account. Google currently labels the sign-in with our Supabase authentication hostname; its consent page links back to the MacroLens privacy policy and terms.

The terminal is open to everyone

Every desk board, the transmission matrix, scenario lab, peer map and sample brief can now be opened without signing in or sharing an email. Accounts are optional for the daily Morning Desk Note. A prominent "Get the daily brief" button opens a past-issue preview with the delivery rhythm and optional signup.

Clearer market-price attribution

The desk now distinguishes official publisher figures from Reuters-attributed market reports when direct exchange settlement data is unavailable. Each reading keeps its date and a link to the report used. The date-cadence health check is not presented as a full numerical audit.

The full terminal is free with an account

Updated 30 Sep: the terminal is now open without an account. Accounts are optional for the daily email.

Every board and model is available after free account creation. No trial, checkout or new membership is needed. The Morning Desk Note is for registered accounts, with one-click unsubscribe on each issue. Earlier entries below describe the previous paid setup.

Subscribe now charges today, trial stays free

Every plan card now leads with a Subscribe now button: create an account, add a card on Stripe's secure checkout and the first month is charged today - no trial days in between. The 3-day no-card free trial stays as the plain link under each plan, unchanged. Two early Individual subscribers are rate-locked at LKR 5,000/month for as long as they stay subscribed.

Try the desk for 3 days without a card

New accounts get full trial access as soon as they sign up. The trial ends after 3 days with no automatic charge; add a card through Stripe only if you choose a paid membership. Existing card-based trials keep the terms they started with.

Show or hide your password

Password fields now have an eye button so you can check what you typed before signing in, creating an account or changing a password. The password stays hidden by default.

Every plan now starts with 3 days free

Desk Note, Individual and Enterprise all open with a 3-day free trial. You add a card through Stripe's secure checkout and nothing is charged for three days. Cancel before the trial ends from the Account page and you pay nothing; otherwise the membership starts after 3 days and renews monthly. Trial members see a countdown on the desk and can start their membership early in one tap. One free trial per account.

New GDP nowcast vintage on the Q2 print, and truer CSE dates

The Q3 2026 nowcast is re-anchored on the official Q2 print (+4.2%, DCS, 15 Sep) and now reads +4.6%, band 3.6-5.6. The 27 Aug vintage (+5.5%) stays on record and both are scored when the Q3 print lands. Peer-map prices are now dated by the last CSE trading day rather than the calendar day, so nothing says "today" before the market opens, and a stock that times out is retried once.

Every dated board checked against its publisher's cadence

The desk now keeps a freshness register for each board: publisher, how often it updates, and the date shown. Boards that had fallen behind were brought current from the latest official release: power mix to CBSL's 18 Sep weekly indicators, the food basket to CBSL's 22 Sep daily price report, reservoirs to the Irrigation Department's 13 Sep reading, tourism to SLTDA's 1-14 Sep weekly report, Hambantota to its record July, and the mobile market total to TRCSL's July statistics. The Brent chart now runs through 22 Sep. Where a publisher has not released anything newer, the board keeps its date and says so.

Live CSE peer prices, a steadier Ask the Desk and cleaner board popups

The listed-peer map now pulls prices from cse.lk on the server, with a date stamp and a health check, instead of a fixed snapshot. Ask the Desk now answers company market-cap questions directly. Board popups open as a single, readable table.

Published vintages separated from new outcomes

The live desk now distinguishes the frozen 27 August Q3 nowcast from the Q2 GDP outcome published on 15 September. Stale GDP, tourism and Treasury calendar copy was corrected, the rupee gauge fallback was reconciled to its published formula, and the methodology now states how the scenario matrix reaches each result.

A plain description of the product

The landing page and link previews now describe MacroLens directly as a live Sri Lankan macroeconomic terminal. The third-party product comparison has been removed from public metadata.

One calmer subscription palette

The three membership cards now use the cooled ivory identity and one disciplined blue-grey accent family. Enterprise keeps a clear visual lead without the competing green, navy and yellow blocks, while every Subscribe button retains strong accessible contrast and the checkout flow is unchanged.

Card payments are live

All three plans - Desk Note, Individual and Enterprise - now check out through Stripe's secure card payment page. Membership activates automatically once payment is confirmed, renews monthly, and can be cancelled from the Account page with access running to the end of the paid month. The manual bank-transfer checkout and the three-day trial are retired; membership starts with payment.

A calmer, single-line header

The terminal and landing headers were rebuilt on a single-line pattern in the spirit of Apple's navigation: a slimmer 52px translucent bar, evenly spaced 12px destinations, and a More menu that can never drop onto a second row. On phones, a compact menu button opens a full-screen destination menu with large readable links, while the section rail stays for one-tap jumps. Search, pricing, plan status and account controls are all preserved, and Escape closes every menu.

Headline numbers lead the desk

The Live desk now leads with the GDP nowcast and headline-number boards, with the rupee and FX kept high in the first scroll. Dated events and the analyst read follow, then commodities, exports, operating-resource and gateway detail. Population and labour remain lower, while corporate wallet and market-share detail now sits last. All boards remain available; only repetition and scroll priority changed.

Port traffic without the broken frame

The current SLPA berthing programme remains the native view for vessel ETAs and planned moves. Live AIS positions now open on verified external maps instead of a third-party frame that refused to load.

A quieter route through the terminal

The desktop header now keeps the core workflow visible - Transmission, Live desk, Forward and Scenario lab - and moves supporting destinations into a More menu. Search, pricing and account controls have clearer space. On smaller screens, the complete section rail remains available.

Procurement joins MacroLens

The new Procurement Desk tracks official tender notices across Treasury, Health, CBSL, CEB and SLPA. Search, category filters, deadline signals and addendum flags are now inside the MacroLens terminal. Every record links to the issuing authority and separates source facts from MacroLens-parsed fields. The opening snapshot contains nine records and makes no all-government coverage claim.

Power, trade, tourism and food

Four full sector desks were added to the terminal: plant-level generation mix, monthly merchandise flows, tourist arrivals with source markets and earnings, and daily food-price movements. Each desk names its official publisher and issue date.

Scenario lab assumptions made explicit

The sensitivity engine was audited across all 11 sliders. Stale baselines were wired to the live prints; the unsupported one-standard-deviation claim was removed; timing, linearity, additivity and overlapping-driver limits are now disclosed beside the controls. The lab opens on its unchanged baseline.

Enterprise data and narrative pass

The data spine was checked against official CSE, CBSL, DCS, SLPA and AASL releases. Thirteen fixes shipped, including stale inputs, narrative conflicts, pricing copy and an incorrectly labelled coconut line. The indicator library and desk read now refresh with the daily pipeline.

Headers tightened and exposed key restricted

Security headers and robots.txt were deployed site-wide. The browser AI key was found in page source and restricted to macrolens.lk, then tested as blocked off-site while the desk chat remained operational. Corporate-filter reclassification requests were also submitted.

17-indicator library and methodology

The Indicator Library now carries 17 live Sri Lankan indicators with latest print, prior change, date, percentile, history and source. The methodology page publishes the nowcast method, rupee-gauge formula and benchmark working, and states plainly where the live model has no scored forecast history yet.