Bali's Subak Irrigation Governance

Water makes a constituency that village lines do not contain

Volcanic slopes and seasonal flows do not distribute irrigation water evenly. Weirs, tunnels, canals, gates, and field channels join upstream and downstream farmers whose planting decisions also change the habitat available to crop pests. A farmer can work a plot, but cannot maintain or schedule the relevant water network alone.

Bali's 2012 provincial regulation defines subak as a traditional organization in water and crop management that is social-agrarian, religious, and economic. It names members, field subdivisions, a leader or pekaseh, customary rules or awig-awig, and the deliberative meeting or paruman; it assigns subak functions in water distribution, cropping, maintenance, dispute resolution, ritual, and member welfare. The same regulation grounds subak in Tri Hita Karana: relations with divinity (parhyangan), among people (pawongan), and with the environment (palemahan).¹

That is a formal legal description, not proof that every subak uses the same rules or that all members exercise equal influence. The regulation leaves membership and detailed governance to each group's awig-awig, while requiring those rules and leaders to be approved or chosen through member processes. It recognizes a layered institution: canals and fields establish material interdependence; assemblies make operating decisions; customary rules preserve commitments; and ritual helps constitute the relationships through which those commitments have meaning.¹

Scheduling joins water management to pest management

Planting simultaneously across nearby terraces can interrupt pest habitat, but simultaneity also concentrates demand for water. Staggering cultivation can reduce peak demand while leaving continuous food for pests. A 2017 study joined satellite-observed synchrony in six rice-growing regions to an evolutionary-game model of that tradeoff. Its simulated subaks converged on groups whose schedules resembled the spatial pattern observed from satellites; the authors also report earlier field measurement in ten subaks in which irrigation flow tracked planting schedules.²

This supports a coordination mechanism, not a claim that ritual alone caused the landscape pattern or that one island-wide hierarchy optimizes Bali. The authors warn that similar spatial patterns can have other causes, and the model produces smaller, locally independent groups rather than universal coordination across every subak on a river. Water temples and federated meetings therefore matter as places where relevant farmers can align schedules; they do not amount to a central control room.²

The history is neither timeless nor purely bottom-up

Balinese inscriptions from the ninth through fourteenth centuries document irrigation works, taxation, cultivation, and relationships among rulers, villages, and religious institutions. They establish a long history of coordinated irrigation, but not an unchanged modern subak extending intact from the ninth century.³

Historical research in the former south-Balinese kingdom of Mengwi finds that precolonial rulers participated in the construction, upkeep, and ritual regulation of large irrigation systems. Dutch colonial government then took over regional management after 1906, strengthened and standardized local associations, and reorganized administrative terms and structures. This Mengwi-based account directly challenges a history in which autonomous farmer groups alone built the system, while its regional scope also cautions against substituting one origin story for all of Bali.⁴

Temple authority could be materially unequal as well as integrative. Palm-leaf records concerning Batur Temple describe landholding, tribute, labor, and religious-economic obligations connecting peasants, temples, and dynasties. That study's late-precolonial Batur setting cannot establish how present-day subaks operate, but it shows why sacred water should not be equated with the absence of hierarchy.⁵

The 2012 regulation does not classify members by caste, and the operating studies reviewed here do not supply a contemporary caste-stratified audit of water shares, meeting voice, or sanctions.¹⁶⁷ Neither fact establishes that caste or social rank is irrelevant in practice. A defensible account keeps local water deliberation distinct from the wider ownership, household, temple, and political relationships in which it occurs.

The Green Revolution tested scheduling authority

In the 1970s, Indonesian agricultural and irrigation programs promoted new rice varieties, fertilizer, pesticides, and more frequent planting while taking planting schedules away from subak assemblies. The 2017 study reconstructs a rapid sequence: Green Revolution rice reached about 70 percent of southern Bali's terraces by 1977; initial yields rose; within two years officials were reporting water-scheduling disorder and pest outbreaks; and farmers later resumed synchronized planting, after which yields recovered. A 1985 Tabanan Public Works report and the Asian Development Bank project's final evaluation are among the underlying institutional records.⁸

The sequence is strong evidence that scheduling authority had operational consequences. It is weaker evidence about distribution and political agency. The published reconstruction does not identify which farmers first resisted, how losses differed among owners and tenants or upstream and downstream fields, or whether restored yields were shared evenly. Its model tests a water-pest mechanism; it does not turn the episode into proof that all decentralized decisions outperform public agricultural programs.⁸

Membership does not make every worker equally heard

A two-year participatory project in the Catur Angga Batukaru part of the World Heritage landscape held individual and group discussions with farmers, pekaseh, government officials, researchers, and civil-society participants. Farmers described rising production and ritual costs, inadequate returns, water and land pressure, and little benefit or decision-making power in tourism. In a survey of 156 farmers across six subaks, 55 percent were at least 50 years old and about 13 percent were 35 or younger. Women in project discussions said they did not want their children confined to poorly paid farming, while also proposing agricultural education and ways for farmers to benefit from the heritage economy.⁶

The project also observed that pekaseh were men and that women who performed substantial field and ritual work commonly voiced concerns indirectly through a male household head, unless they themselves represented the household. A 2023 critical ethnography in Subak Bulung Daya similarly describes women's farm and ceremonial labor alongside marginalization and a double burden.⁷ Neither study measures prevalence across Bali: the first is a participatory project concentrated in one heritage component, and the second is a single subak case. Together they make a narrower point that household membership and collective ritual can coexist with gendered labor and voice.

Tenants, sharecroppers, hired workers, absent owners, and households with very small holdings can also carry different obligations and exposure, but the reviewed studies do not provide a comparable Bali-wide accounting of their water, votes, income, and losses. “Farmer-governed” identifies an important location of authority; it should not erase differences among farmers or between land ownership and agricultural work.

Heritage protects selected landscapes, not the whole institution

In 2012 the World Heritage Committee inscribed a serial property made up of five selected landscape components. Its decision recognized an officially defined cultural landscape of water temples, terraces, canals, forests, villages, and Tri Hita Karana. It did not inscribe every subak in Bali, and the designation's statement of value is not an evaluation showing that inscription improved farm income, representation, water security, or ecological condition.⁹ The UNESCO World Heritage record is an authoritative destination for the property's boundaries, documents, and formal status.

An ICOMOS-ICCROM advisory mission in 2015 found no direct subak representative in the property governing assembly and recommended direct participation, financial support, coordinated land-conversion control, tourism planning, and catchment management. The mission was a short, invited advisory review rather than an independent longitudinal evaluation, but it identifies a structural problem: an institution can be central to a protected landscape without its members controlling the protection regime.¹⁰

The Catur Angga Batukaru project found a related gap from below. Participating farmers often described themselves as spectators to tourism, asked for clearer information and representation, and proposed markets, training, rules, and incentives that would connect visitors' spending to continued farming. Those are situated participant accounts, not evidence that every heritage-area farmer shares the same view.⁶

Tourism and land conversion reach beyond awig-awig

The provincial regulation formally requires member agreement and regent or mayor approval for conversion of subak agricultural land. It also requires any outside use of subak irrigation water to be coordinated with all members through the assembly. Those provisions recognize collective stakes, but the text alone does not show enforcement or give one subak control over groundwater licensing, regional infrastructure, property markets, and tourism planning.¹¹

A social-ecological study of tourism and water inequity reanalyzed interviews, focus groups, and visitor surveys conducted in 2010. It describes stratified water users, fragmented government responsibility, strong tourism-industry influence, and little subak participation in decisions beyond local irrigation. The community interviews and pekaseh focus group were concentrated in Canggu, and the authors call their systems analysis preliminary; it establishes a plausible governance mechanism and local experience, not a current island-wide measure of water capture.¹²

Land conversion is observable in bounded urban cases. A 2024 spatial study of four South Denpasar subaks compared mapped area and inspected irrigation networks, reporting a reduction from 292.593 hectares to 186.314 hectares and the disintegration of Subak Sesetan. Its GIS and field observations document one urban district, not the rate or cause of conversion across Bali.¹³

The comparison with Japan's iriai forest commons is therefore useful but limited. Both join membership, place, customary obligation, and a shared resource, and both can be weakened by transactions that reward one owner while fragmenting a collective system. Subak adds an especially visible dependence among cropping schedules, canal flows, and ritual authority. In both cases, customary rules need public law and material livelihoods that protect collective action without replacing member voice. The relationship is an analytical comparison, not a claim of direct institutional influence.

Current protection is a warning, not a conservation verdict

In July 2025 the World Heritage Committee welcomed progress on spatial planning but expressed concern about tourism and development, particularly at Jatiluwih. Its adopted decision said existing measures had been insufficient to stem overtourism and encroachment, warned of rice-field loss and fragmentation and weakened subak authority, and described the property as approaching a “tipping point.” It requested stronger permit and land-sale enforcement, support for the socioeconomic and cultural conditions of farming, detailed land-use maps, an advisory mission, and another state-of-conservation report by 1 December 2026.¹⁴

That decision is current formal monitoring, not a measured estimate of how much land, water, income, or authority every subak has lost. Its requested reporting deadline was still ahead on 14 July 2026.¹⁴ The justified conclusion is that the international governing body considered protection inadequate and sought more data and action—not that World Heritage status had either saved or destroyed the institution.

Ecology is more than rice yield

The water-pest study explains how coordinated planting can change hydrological demand and interrupt pest habitat.² A separate 2021 study in Jatiluwih used purposive vegetation plots, drone mapping, and key-informant interviews to document plants used for food, fodder, ceremony, medicine, construction, and craft across rice fields and adjacent land; the non-rice tegalan plots had the highest measured diversity.¹⁵ This is evidence that one managed subak landscape contains useful plant diversity, not that every subak increases biodiversity, avoids pesticides, or benefits every nonhuman species.

Claims about “balance with nature” should remain testable rather than ceremonial. Useful measures would distinguish surface water from groundwater; record water quantity and quality upstream and downstream; map land conversion and canal breaks; track pesticide and fertilizer use; and observe pest, soil, plant, and aquatic communities over time. Ecological outcomes cannot be inferred from formal Tri Hita Karana language alone.

The organizational fingerprint distinguishes coordination from romance

Purpose, mission, and institutional legitimacy is defining because water distribution, viable cultivation, canal maintenance, ritual obligation, and member welfare are joined rather than treated as separate programs in provincial law. Historical and ethnographic evidence shows that the institution has changed, so legitimacy cannot rest on a claim of timelessness. It must also come from whether participants can continue farming, deliberate, and sustain the relationships that give the rules meaning.¹³

Authority, legitimacy, and acceptance is defining. Water flows and field interdependence create material authority; assemblies, leaders, and awig-awig create customary authority; temples and ritual create sacred authority; provincial law and heritage bodies add state and international authority. Those sources can reinforce or contradict one another. Formal recognition of a subak does not prove equal acceptance of a decision by upstream and downstream farmers, women, tenants, or households facing land-sale pressure.¹¹⁰

Delegation, decentralization, and responsibility scores strongly. Members and a pekaseh handle local distribution, maintenance, schedules, disputes, and ritual, while federated temple meetings can coordinate a larger hydraulic area. Wider groundwater, tourism, land-use, and heritage decisions remain with public and market actors. The 2015 advisory mission's finding that direct subak representation was absent from the property governing assembly shows how responsibility for maintaining a landscape can be decentralized while development authority is not.¹¹⁰

Coordination, communication, and common understanding is defining because one farmer's planting date affects neighbors' water and pest conditions. Assemblies, customary rules, shared maintenance, and temple networks make that interdependence discussable. The water-pest study supports a mechanism through which local schedules can produce larger spatial patterns; it does not show that every participant understands the model or that ritual eliminates conflict.²

Structure, hierarchy, and scale scores strongly. Field channels, subaks, temple relations, watersheds, districts, provincial government, and the World Heritage serial property operate at different boundaries. Subak is neither one island-wide hierarchy nor a set of isolated villages. Historical evidence from Mengwi and Batur also shows that rulers, temples, and colonial administrators helped shape regional structure, qualifying a purely bottom-up account.⁴⁵

Decision making, judgment, and bounded rationality scores strongly. Farmers decide under uncertain rain, water, pest, labor, price, and ritual conditions, and no single schedule maximizes every objective. Local knowledge helps interpret a gate, field, or outbreak; coordination makes wider effects visible. The model clarifies the water-pest tradeoff but cannot select a just allocation or determine whose loss should count most.²

Measurement, accounting, and control scores lightly. Water turns, crop calendars, land area, contributions, harvest, and sanctions can all be recorded or observed, while ritual and reciprocal work also coordinate without a single ledger. Recent spatial mapping and heritage monitoring make land conversion visible, but bounded maps do not supply an all-Bali account of water, tenure, income, participation, or ecological effect.¹³¹⁴

Cooperation, incentives, and organizational equilibrium is defining. Upstream restraint, synchronized or staggered planting, canal labor, dispute settlement, and ritual can preserve a shared production system whose benefits no plot can secure alone. Cooperation becomes fragile when farm returns fall, tourism captures value, one land sale fragments canals, or wider water decisions ignore subak. The participant project documents those pressures in one heritage component rather than an island-wide equilibrium estimate.⁶

Governance, stewardship, and accountability is defining because water and landscape care require rules, recognized decision makers, review, and consequences. Provincial law assigns functions and formal conversion procedures, while heritage monitoring asks public bodies to control encroachment and include farmers. Neither source demonstrates compliance. A complete accountability chain would trace a proposed water change or land sale through member deliberation, public permit, physical result, remedy, and later ecological condition.¹¹¹⁴

Culture, informal organization, trust, and voice scores strongly because sacred practice, household representation, reciprocal labor, and expectations surrounding meetings shape what formal rules can do. Gender-focused evidence in one heritage component and one subak identifies substantial women's work and constrained direct voice. It establishes a serious mechanism and situated experience, not a prevalence rate across Bali or proof that all households organize representation alike.⁷

Knowledge, expertise, and professional autonomy scores strongly. Farmers possess place-based knowledge of gates, soil, crops, pests, and obligations; priests and leaders hold ritual and institutional knowledge; hydrologists, planners, conservation specialists, and tourism bodies bring other expertise. Heritage protection can help assemble knowledge while also displacing farmers from interpretation. The record supports participation as a design need more securely than a claim that every expertise conflict is resolved locally.¹⁰⁶

Learning, quality, and reliability is defining. Seasonal repetition lets farmers observe water, pest, and harvest consequences and adjust schedules; the Green Revolution episode provides historical evidence that external scheduling disrupted those feedbacks and that coordination later returned. The reconstruction cannot disaggregate every actor or prove that restored local authority alone caused yield recovery, but it shows that learning resided in the coupled operating system.⁸

Work design, productivity, and automation scores zero as an emphasis. Subak coordinates field and maintenance labor and can affect yield reliability, yet the reviewed evidence does not show that redesign of jobs, automation, or measured labor productivity is the institution's distinctive mechanism. Farmer accounts instead expose poorly paid work, ritual costs, aging membership, and gendered burdens that headline yield figures can hide.⁶⁷

Strategy, competition, and adaptation scores lightly. Subaks adapt schedules and rules and now face competition from tourism and land markets, but they do not primarily pursue competitive strategy. Their strategic problem is maintaining the material and legal conditions under which coordination remains possible when actors outside the water network can profit from conversion or extraction.¹²¹⁴

Innovation, entrepreneurship, and renewal also scores lightly. Long-lived institutions change: colonial administration, agricultural policy, provincial law, heritage designation, and farmer proposals have all recombined practices and authority. Current proposals to connect tourism revenue or education to farming may support renewal, but the cited participatory project records proposals rather than evaluated innovations at scale.⁶

Executive attention, information, and organizational sensing scores zero. Leaders and heritage bodies receive reports, and the 2025 World Heritage decision is an important alarm about fragmentation and overtourism. The evidence does not show a distinctive executive-sensing architecture inside subak or a closed loop from alarm to enforcement and restored farming conditions.¹⁴

Organizational ignorance also scores zero as an emphasis. The record contains explicit unknowns about membership, water shares, tenure, gender, conversion, income, and ecology, but it does not show an institutional practice devoted to finding and governing those unknowns. Formal Tri Hita Karana language can itself conceal gaps if cultural, social, and environmental harmony are asserted rather than measured.¹⁶

The relationship to benefit for all life is an editorial lens, not historical influence. Subak directly joins human livelihood, water, crops, pests, plants, ritual, and future continuity, making a multi-subject review appropriate. The evidence supports mixed and bounded effects; it does not justify an aggregate claim that the institution benefits every member, species, ecosystem, or future generation.¹⁵¹⁶

What remains unresolved

The available record is unusually strong on formal rules, the water-pest mechanism, selected historical transformations, and the monitored World Heritage property. It remains fragmented on distribution. No reviewed source provides a current, reconciled all-Bali account of membership and land tenure; water shares and sanctions; participation by gender, caste, age, and work status; tourism revenue reaching farmers; conversion and canal loss; or linked ecological outcomes. Provincial rules describe authority but not compliance. Historical and ethnographic cases reveal forms of power but do not measure their current prevalence. Models isolate mechanisms but do not settle political causation.

The clearest present claim is correspondingly bounded. Subak gives many farmers a durable institution for coordinating irrigation, cultivation, maintenance, and ritual across interdependent fields. That local and federated capacity has survived major interventions and remains materially consequential. Whether it can continue depends not only on internal cooperation, but on who owns and farms the land, who speaks in households and assemblies, who controls wider water and development decisions, and whether farming remains a viable way to inhabit the landscape.

Paths for deeper study

• Compare awig-awig, meeting records, water schedules, sanctions, and dispute files across upstream and downstream subaks, including groups outside the World Heritage property.
• Pair owner and leader interviews with women farmers, tenants, sharecroppers, hired workers, young people, temple personnel, and households that sold or stopped farming; publish how participants controlled attribution and use.
• Reconcile cadastral, irrigation, permit, tourism, and water-abstraction data with field mapping, and separate formal approval from actual conversion.
• Track farm income, tourism revenue, ritual and maintenance costs, water reliability, crop loss, chemical use, and biodiversity with methods that can distinguish subak effects from rainfall, urbanization, market change, and public infrastructure.
• Follow the state-of-conservation process requested for December 2026 and test whether new planning and enforcement alter land conversion, farmer voice, and livelihood conditions.