Recorded Alcoholic Beverages in Sri Lanka: Supply, Excise and the Limits of the Market Evidence

Recorded supply recovered; revenue measures something different

Sri Lanka’s recorded alcohol system expanded in 2025, but its production and revenue indicators moved at different rates. Treasury’s Final Budget Position Report 2025, Table 4.8, reports 37.3 million absolute litres of hard and malt liquor production, against 33.4 million in 2024. Net liquor receipts rose 8.51%, more slowly than the 11.7% production increase. These are useful measures of the recorded system. Neither is a retail market valuation or a count of what residents drank.

The distinction changes the interpretation of recovery. Production can be exported or enter stocks; tax is affected by rates, exemptions, timing and accounting. Higher receipts alone establish neither higher consumption nor better compliance. This audit follows the quantities and money separately, then identifies the missing connections.

Evidence checked through: 20 October 2026. Coverage is limited: recorded production, fiscal accounts, operative tax distinctions, regulatory controls and selected participant activities are examined. A complete domestic-release balance, national retail value, company shares and representative margins remain unavailable.

The boundary follows products and transactions

The market examined is licensed and duty-recorded alcoholic beverages released for Sri Lankan domestic consumption. Imports, distribution, licensed retail and hospitality sales belong to that domestic boundary. Manufacturing and export records provide separate supply context; unrecorded production is an external alternative requiring its own measurement. Industrial alcohol and non-alcoholic beverages are excluded from alcohol volume totals.

An absolute litre measures ethanol content. Finished-beverage litres, bottles and absolute litres cannot be added. Manufacturing output, domestic duty-paid releases, cash excise receipts, issuer revenue and household spending are also different stages. The analytical bridge would be production plus imports, less exports and changes in stocks, aligned with taxable releases and exemptions. The inspected records do not complete that bridge. A residual is consequently left unexplained rather than assigned to illicit activity.

The intended use is to assess recorded supply, fiscal exposure and regulatory conditions. It does not support forecasts of consumer demand, valuations of producers or commercial targeting based on health-survey segments.

Product-specific production shows an uneven recovery

Source: Treasury, Final Budget Position Report 2025, printed p.86; its accompanying figure labels 2025 provisional. Changes are (later / earlier − 1) × 100, using the displayed rounded quantities. Hard liquor accounted for most of the increase, while malt production also rose. This is a production comparison, not a consumer preference study. The source changes alcohol-volume units after 2017; pre-2018 figures are not spliced into this series. Wine, toddy, other categories and imported finished products lack a complete comparable release account here.

There is material counterevidence to a seamless official series. The Excise Department’s 2024 report says liquor production increased 1.4% in 2024, whereas the later Treasury hard-plus-malt series falls from 34.0 to 33.4 million absolute litres. No explicit scope or revision bridge was located. Both statements remain recorded; the table uses the identified Treasury vintage consistently.

The 2026 Budget report, printed p.30, has a separate January–August 2025 total of 23.9 million absolute litres, but its hard and malt components sum to 23.7 million. That 0.2 million residual is not silently removed or interpreted as unrecorded supply. Partial-year figures are not substituted for the annual series.

Net receipts, export relief and statutory rates need separate accounts

The net amounts are Treasury revenue code 1002.04.01; the 2024 number also matches the Excise financial table. Adding the Auditor General’s section 1.6.1 export-relief amount reproduces the gross reconstruction. This preserves the earlier correction: the auditor found that netting relief understated both revenue and expenditure. The Accounting Officer explained that the existing method and note disclosure had been retained because of difficulties making the repayment through revenue or an expenditure head. The reconstruction does not declare that accounting dispute resolved.

The Excise narrative’s rounded Rs 225.5 billion liquor figure and Rs 226.7 billion including tobacco must not replace its net financial table. Nor should its reported 27% revenue growth be compared directly with growth in Treasury’s net series without a gross-to-net bridge.

Excise Notification 01/2025, dated 10 January and effective 11 January 2025, establishes distinct duties per litre of alcohol. Selected domestic categories are special arrack at Rs 7,244, specified other arracks at Rs 7,752, country-made foreign spirits at Rs 7,969, malt liquor at 5% strength or below at Rs 5,735, and malt liquor above 5% at Rs 6,015. These are not duties per litre of finished drink. Clauses 10–11 separately address exports and specified diplomatic supply.

The operative text matters: Treasury’s later Table 4.9 places exactly 5% beer in the higher band, contrary to that notification. The notification governs this comparison. The Budget report’s narrative also says 1 January, while its measures table and the notification specify 11 January. These source errors are retained in the research record.

The January–August 2026 Fiscal Review reports liquor receipts of Rs 181.1 billion, against Rs 145.7 billion in the same 2025 period: 24.30% from rounded inputs. This is a later fiscal signal, not a forecast for the full year or proof of a similar increase in production. Department-of-Excise totals include other tax categories and are not an alternative liquor series.

Participants compete through a regulated value chain

The observable sequence is inputs and packaging, licensed production or import, storage and transport, authorised distribution, then retail or hospitality service. Taxable releases and export movements require separate records. Licence conditions and lawful access constrain each route; a brand portfolio alone does not demonstrate route coverage or bargaining power.

Two current issuer examples make the activity boundary concrete. Lion Brewery’s 2025/26 report, covering the year to 31 March 2026, describes brewing, packaging, distribution and sales, with local and export business. Its value-chain account distinguishes off-trade shops from pubs, bars, hotels, restaurants and clubs, and describes third-party local transport. It also includes contract brewing in Vietnam. Those overseas quantities cannot automatically be counted as Sri Lankan production.

Distilleries Company’s 2025/26 report provides a second first-party observation of liquor manufacture and distribution. These examples establish participant activities, not a complete competitor census. Imports, other licensed producers and unrecorded alternatives remain part of the competitive question. No national shares, concentration index, retailer margins or representative producer profitability are calculated from the two companies.

Interpretation: measured alcohol strength, packaging availability, lawful release records and continuity of distribution are economically relevant because they determine what product can move through the regulated chain. The strength of supplier or buyer bargaining power is not measured by the available evidence. A full assessment needs current supplier dependencies, transaction terms, licence status and channel-level economics.

Demand evidence describes exposure, not a current sales basket

The STEPS 2021 fact sheet estimated past-30-day alcohol use at 20.7% among adults aged 18–69: 43.3% among men and 1.2% among women. Heavy episodic drinking was 7.1% overall under the survey’s six-or-more-drinks definition. Fieldwork occurred in April, November and December 2021 and was interrupted by the pandemic. Self-report and population coverage limit interpretation.

The user, household payer and licensed hospitality buyer need not be the same decision-maker. Health-survey sex differences help identify exposure and service needs; they do not establish brand preferences, spending, price elasticity or a commercial segment size. A current matched price, income and purchasing panel was not found. Consequently, substitution after a tax change remains a hypothesis requiring evidence.

WHO’s 2024 global report retains a different historical benchmark: modelled 2019 total consumption of 2.8 litres of pure alcohol per person aged 15+, with a 1.4–4.4 uncertainty interval. Its rounded recorded and unrecorded components, 2.6 and 0.3 litres, do not sum exactly to the displayed total. The modelled 3,128 attributable deaths and 153,977 disability-adjusted life years are not current observed case counts.

A March 2026 parliamentary record attributes a 30–35% illegal-consumption estimate to an Accounting Officer citing institutional studies. It does not supply their methods. This cannot be combined with WHO estimates or the fiscal residuals to derive a national illicit share.

Authority, access and control failures shape the outlook

The NATA Act gives public-health oversight a different purpose from revenue collection. Section 31 restricts sales to people under 21, and sections 35–37 regulate advertising, sponsorship and promotions, with specified exceptions. Excise administers licensed production and sale; Treasury determines fiscal policy; Customs deals with imports; the Auditor General and Parliament provide scrutiny. Consumers, workers and communities bear exposures that fiscal totals do not measure.

The 2024 audit identified 88 RB.04 licences issued during the presidential-election period without the formal process urged by the Election Commission. That is a finding about a defined set of decisions, not a count of all outlets. It supports treating administrative transparency as a material access condition rather than assuming that licensing is uniformly applied.

Traceability technology is also a testable control, not a claim of success. The audit distinguished 54,887 bottles suspected of fake stickers in selected Northern and Eastern inspections from 12,170 identified as carrying fake stickers, and described 8,446 illegally manufactured bottles bearing official stickers. These targeted observations cannot estimate national prevalence. Implementation should be judged through matched production, labels, releases and exceptions, including how discrepancies were resolved. Product novelty, automation expenditure or a secure label by itself does not establish compliance, quality or lower harm.

The useful outlook is conditional

As assessed on 20 October 2026, this outlook covers the period to 20 April 2027, with earlier reassessment on material rule or evidence changes.

The evidence supports monitoring continued recorded-production recovery and rising receipts, with uncertainty about how much reaches domestic consumers. There is no probability-weighted sales forecast. If future releases rise alongside production and inventories remain stable, domestic-supply growth becomes more credible. If output rises mainly through exports or stocks, the domestic-demand interpretation weakens. A tax-rate or accounting change can invalidate a simple receipts comparison even when physical supply is unchanged.

For lawful participants, three external requirements follow from the findings. Product and tax classification requires records linking strength, category and release to the correct duty; the observable test is a reconciled product-level declaration. Distribution integrity requires traceable stock and valid permissions across producer, transporter and outlet; the test is a resolved exception trail rather than sticker counts alone. Control over fiscal presentation requires gross duties, relief, arrears and cash receipts to be separately explainable; the test is a reproducible bridge across accounts. These requirements derive from the tax, channel and audit evidence and are not claims that any named company already meets them. Changed rules or independently verified resolution of the identified failures would require reassessment.

There is a public-interest opportunity in clearer release data and auditable administration. Threats include opaque licensing, unresolved label exceptions and dependence on fiscal aggregates that obscure product changes. The record does not justify promoting alcohol consumption, forecasting producer returns or treating health controls as a commercial loss by definition.

Coverage, comparability and the next evidence checks

The fiscal and production baseline is stronger than the retail, channel and customer evidence. Missing releases, imports, exports and stock movements prevent a complete supply balance; missing comparable prices and costs prevent retail valuation and profitability estimates. Licence counts do not equal active firms, and two issuer examples do not represent all producers. Indian or Vietnamese growth rates cannot fill those gaps or validate a Sri Lankan forecast.

The 2025 Excise performance report was presented to Parliament on 9 July 2026, but its underlying PDF was not located in the reviewed archive paths. Its retrieval, subsequent fiscal reviews, new duty notifications, updated licence records and audit responses are continuing source obligations. New STEPS or WHO releases should update health findings only after checking populations, definitions and observation periods. A later source does not erase an earlier contradiction without explaining the revision.

This edition retains the original 23 June 2025 publication date and the 14 August 2026 correction. Unsupported market forecasts, shares, numerical competitive scores and company investment programmes removed by that correction remain excluded. The source dates distinguish historical observations from records inspected on 20 October 2026; the evidence does not become a 2026 observation merely because it was accessed then.