What Permendag 19/2026 Does
Indonesia's Ministry of Trade signed Regulation No. 19 of 2026 (Permendag 19/2026) on June 4, 2026, entering it into force on June 8. It replaces Permendag 31/2023 — the framework that first imposed systematic obligations on e-commerce platforms — and extends the PMSE (Perdagangan Melalui Sistem Elektronik) licensing regime to eight business models for the first time, including ride-hailing apps and online travel agents. Trade Minister Budi Santoso framed the overhaul as an effort to "strengthen a digital trade ecosystem that is fair, healthy, and beneficial."
The timing matters. Indonesia's e-commerce GMV exceeded $75 billion in 2024 and the country accounts for roughly 44% of Southeast Asia's entire e-commerce market. With 212 million Indonesians online as of early 2025, the scale of what Permendag 19/2026 governs is not abstract — it shapes the operating conditions for virtually every major consumer-facing platform in the fourth-most-populous country on earth.
Where the Regulation Gets It Right
The strongest provisions are those grounded in consumer transparency rather than market engineering. Article 47(3)(a) requires businesses using AI to "disclose to consumers when products, services, or recommendations are generated, displayed, or promoted using AI." That obligation is narrowly drawn and proportionate to a real consumer protection concern: AI-driven recommendation systems that behave like editorial choices but are presented as neutral ranking. Platforms must additionally maintain internal AI governance proportionate to risk and provide correction mechanisms — norms broadly consistent with what the EU's Digital Services Act already requires of algorithmic recommender systems, and with practices major global operators have already adopted.
The KPPU reporting mandate is similarly defensible in principle. Platforms must notify Indonesia's competition authority (Komisi Pengawas Persaingan Usaha) within three business days of discovering suspected price manipulation, below-cost selling, or market-distorting subsidies. KPPU has significantly ramped up its digital-sector enforcement in recent years, imposing its highest-ever administrative fine in 2025 and advancing a draft competition law amendment — now in Parliament's 2026 priority legislative programme — that would add extraterritorial jurisdiction and a cartel leniency programme. Requiring platforms to surface suspected violations rather than passively avoid them is a reasonable evolution of that enforcement posture, provided KPPU has the institutional capacity to process disclosures responsibly. The three-business-day window is aggressive but workable for platforms whose trust-and-safety teams already monitor pricing anomalies continuously.
The Algorithm Mandate Problem
The more interventionist provisions concern how platforms must rank and surface products. Permendag 19/2026 requires platforms to configure search results, recommendation engines, and promotional features to prioritise domestic goods, with particular emphasis on micro and small enterprise (MSME) visibility. Platforms must create dedicated promotional pages for domestic products and ensure domestic sellers receive fee incentives.
The case for this is sincere. Indonesian MSMEs represent the backbone of the country's employment base, and cross-border sellers — particularly from China — have achieved dominant positions in several product categories, partly through logistics and subsidy advantages unavailable to domestic sellers. Platforms that aggregate demand while importing supply at below-market cost do generate a legitimate policy concern that goes beyond ordinary protectionism.
But mandating specific algorithmic outcomes is a different and more problematic intervention than mandating algorithmic transparency. Search results configured by regulatory instruction to favour domestic goods are, by definition, less accurate as signals of what consumers want and what sellers merit on quality, price, and service. The regulation avoids specifying algorithmic formulas — requiring only that domestic visibility be "strengthened" — but that ambiguity will be resolved in enforcement, and enforcement-driven algorithm design creates precedents that are difficult to contain. When India attempted something structurally similar under its 2021 e-commerce framework — pushing platforms to prioritise domestic sellers without precise specifications — it generated compliance uncertainty that persisted for years and created genuine deterrents to platform investment. Indonesia, with its $104 billion market forecast for 2026 and broader digital economy ambitions, should be watching that experience closely.
The Ride-Hailing Overreach
The regulation's most contestable boundary is its inclusion of ride-hailing platforms and online travel agents under the PMSE framework. The logic is pragmatic: Gojek and Grab already facilitate the sale of food, groceries, and physical goods through in-app commerce features, and drawing a clean separation between those commerce functions and the rest of the platform is genuinely difficult.
Yet industry groups were right to flag the structural mismatch before the final text was confirmed. Ride-hailing operates through labor intermediation and transportation service contracts governed primarily by transport law, not trade law. Applying PMSE licensing obligations, seller verification requirements, and MSME ranking mandates designed for product marketplaces to mobility platforms creates compliance frameworks that map poorly onto the underlying business model. Permendag 19/2026 addresses this partly by limiting ride-hailing obligations to commerce features within the app — transportation services are explicitly excluded — but that line will be litigated by platforms whose commerce and mobility offerings are deliberately integrated rather than cleanly separated.
Thresholds and the Local Representative Question
For foreign platforms, the KP3A local representative requirement triggers at 1,000 consumer transactions per year, 1,000 packages shipped, or reaching 1% of Indonesian internet users. With 212 million Indonesians online, 1% equals approximately 2.12 million users — a threshold that Shopee (52% market share), TikTok Shop ($6.2 billion GMV in 2024), and other major platforms clear easily. This creates genuine governance obligations rather than a theoretical backstop, which is the point: foreign operators with material market presence will need Indonesian representatives who can interface with regulators and courts.
The threshold design is broadly sound. What remains less clear is whether Indonesia's regulatory institutions — the Trade Ministry, KPPU, and the licensing bodies under the Online Single Submission system — have the capacity to enforce consistently across the full range of newly covered platform categories. Permendag 19/2026 places significant institutional trust in regulators' ability to assess algorithmic compliance claims and adjudicate AI governance proportionality assessments. Building that capacity should accompany, not follow, the regulation's ambitions.
The Bottom Line
Permendag 19/2026 is a serious and substantive instrument. Its AI transparency and KPPU reporting provisions reflect international best practice adapted to Indonesian conditions and deserve adoption as a model for the region. Its algorithmic prioritisation mandates and ride-hailing scope reveal the characteristic tension in Indonesian digital policy: a government that has correctly identified the tools it wants to use but is still developing the frameworks for knowing when those tools become counterproductive. With Indonesia's e-commerce sector forecast to reach $212 billion by 2031, getting that calibration right matters — not just for Indonesian consumers and SMEs, but for every global platform making long-term investment and market-commitment decisions.