Part I — Situation overview

The water management crisis entered its institutional phase on 3 August 2026: according to Portfolio’s report, Tanács Zoltán — with the title used by the press — announced that the government is setting up a new, integrated water science centre. The justification of the announcement points back to a decision taken in 2012. The Water Management Research Institute, founded in 1952, for decades gathered into a single organisation the continuous measurements, research, modelling, flood forecasting and the preparation of government decisions; its funding was withdrawn by the government of the day in 2012, and the institution was then closed. According to the justification, strategic water resource research and the operation of national observation systems did not produce direct market revenue — the announcement, however, states precisely that its social benefit was inestimable. After the closure the professional tasks and the decades of accumulated data assets were fragmented: the operational hydrological observations were taken over by the National Directorate of Water Management, the research, ecological and educational tasks went to various university faculties and to the HUN-REN research network, and in 2022 a fixed-term national laboratory was launched to bring the field together. Today there is no comprehensive state research institute that would handle hydrological, water quality and hydraulic engineering questions as a single organisation.

The data of the day of the announcement add the urgency. According to the communication of the South Transdanubian Water Directorate, the lowest water level ever measured has been broken on the Danube and on the Drava alike: at Dunaszekcső the water level is 52 centimetres and at Mohács 57 centimetres below the record of 2018, and the flow of the river is 750 cubic metres per second. The water level of the Danube is −137 centimetres at Paks, −38 at Esztergom and −24 at Baja; 138 million cubic metres are missing from the water reserve of Lake Balaton and 23 million from that of Lake Velence. In answer to HVG’s question the water service stated that if the 10–20 millimetres of precipitation expected for Thursday arrives on the Austrian catchments of the Inn, the Traun and the Enns, the Danube may rise by as much as several tens of centimetres by the end of the week. At the same time it emphasised that over such a time horizon the meteorological forecast is uncertain, so the rise in water level cannot be guaranteed. Meanwhile in two large Hungarian cities — among them Debrecen — a first-degree water restriction is in force, extending to the watering of gardens, the filling of pools, car washing and the operation of fountains; the mayor of Siófok ordered working from home for municipal employees, and Kaposvár is building two new water production wells and a 450-sensor leak forecasting system for HUF 70 million.

According to MIAK’s reading the announcement is substantive and timely, but the policy question is not whether such an institute is needed. The past fourteen years have shown that the existence of a research institute does not in itself protect either the data assets or professional continuity: the 2012 closure was possible precisely because the institution’s value did not appear in any measurable, public performance. The durability of the new centre will therefore be settled not by the fact of its creation but by its mandate — by what it is obliged to measure, what it is obliged to publish, and into whose decision it is compulsorily plugged. The stadium watering affair that came to light on the same day — where the clubs invoked a drilled well and rainwater, even though no rain had fallen for weeks — shows this same gap from the other side: compliance with the restriction cannot be checked today, because there is no single actor keeping the water abstraction data in one place, publicly.

Part II — Literature foundation

Two conceptual frames give the basis of principle for the proposals. The volume Thinking, Fast and Slow by Daniel Kahneman (Israeli-American psychologist, a researcher of the biases of decision-making, laureate of the Nobel memorial prize in economics in 2002) contrasts two approaches to forecasting — the inside and the outside view — and introduces the concept of the planning fallacy for cases where the estimate is unrealistically close to the best case; more important for today’s subject is the author’s thesis about organisations, according to which organisations are better than individuals at avoiding errors precisely because they think more slowly, and are able to make procedures, checklists, reference class forecasting and premortem analysis compulsory. The work Why Nations Fail by Daron Acemoglu and James A. Robinson (economists, leading authors of institutional economics; they received the Nobel memorial prize in economics in 2024) gives the other side: by their definition an economic institutional system is inclusive if, alongside the security of private property and the impartiality of the legal system, it also gives prominence to the provision of public services, and thereby creates a level playing field for the participants — this yardstick is directly applicable to the publicity of water data and to the equal enforcement of restrictions. The detailed treatment of the literature — author by author, with quotations — can be found in the 6.4 Literature in detail section.

📖 Source: Daniel Kahneman: Thinking, Fast and Slow; Daron Acemoglu – James A. Robinson: Why Nations Fail

Part III — MIAK’s concrete proposal

MIAK proposes three measurable elements to be written into the founding decision. All three cost the least at the moment of the organisation’s foundation, and are the most expensive to make good afterwards.

3.1 A compulsory data publication annex in the founding legislation (at the moment of creation)

The durability of an institution is protected not by its budget line but by its public usability: what the country uses daily is hard to close. MIAK therefore proposes that the founding legislation should have an itemised annex on what the centre is obliged to publish — daily water level and water temperature time series in machine-readable form, a register of the water abstraction permits in force with quantitative data, an up-to-date list of the municipalities under restriction and of the restriction levels, and a unified, national drought index. Publication should not be a decision of the management but a statutory obligation, with a fixed delay and an unchanging format. This proposal builds on the public interface of the K3 pollution monitoring and on the machine-readability requirement of the D2 open data programme. The lesson of 2012 is direct here: an institution that has no publication obligation becomes invisible within a few years — and closing an invisible institution costs nobody anything politically.

3.2 A compulsory, public expert opinion for large water abstraction permits and a public forecast hit rate (from the next permitting cycle)

The centre becomes decision-relevant when it speaks at some point in the official procedure. MIAK proposes that in the case of a water abstraction permit application above a defined quantitative threshold the acting authority should be obliged to request the centre’s written expert opinion, that the opinion should be public, and that if the official decision departs from it, the departure should have to be justified. This is not a transfer of competence: the decision remains with the acting authority, only the professional position becomes visible and unavoidable. The second half of the proposal is the condition of institutional credibility: the centre should also publish the subsequent hit rate of its own forecasts — according to Kahneman’s reference class logic (see 6.4.1) the accuracy of a forecasting organisation improves only if it systematically compares its earlier estimates with the factual data. The water service’s present, expressly correct formulation — that a precipitation forecast for several days ahead is uncertain, and therefore a rise in water level cannot be guaranteed — shows precisely this culture; the proposal adds only that this uncertainty should be measurable and visible retrospectively too.

3.3 A no-duplication test and the consolidation of the fragmented data assets (within 90 days)

The greatest risk is not that there will not be enough money for the new centre, but that it will be a fifth player alongside the existing four. MIAK asks that the foundation be preceded — or at the latest followed within 90 days — by a public task map: who measures what and who models what today among the National Directorate of Water Management, the HUN-REN research network, the university faculties concerned and the fixed-term national laboratory launched in 2022, and which of these is not performed by any player. The new centre should take over only what either nobody has or is running in parallel in several places. The same map should also settle the question of the data assets: the scattered time series should go into a single catalogue, even if physical custody remains in its present place. This step follows the logic of the KI3 measurable bureaucracy reduction, and answers the question that the 2012 decision never posed: the problem was not that an institute is expensive, but that nobody wrote down what would cease along with it.

The three proposals can be strung on a single principle: let the institution be strong not from the announcement but from its obligations. A research institute that has a daily publication obligation, a compulsory role in the permitting procedure and a public hit rate is hard to abolish for three mutually independent reasons — not because it is protected, but because it is used. This is the difference between the institute closed in 2012 and a durable institution, and this difference is settled by the text of the founding decision, not by the intention.

Part IV — Expected effects and risks

Dimension Expected effect Risk
Environment and water management Hydrological, water quality and hydraulic engineering knowledge comes together in one place; the professional grounding of permitting improves If the mandate narrows to research and it has no procedural role, the centre remains a producer of studies without any effect on decisions
Public administration The task map clarifies today’s fragmented responsibilities; public data publication reduces the burden of data requests A new player alongside the existing ones, with clashes of competence and the withdrawal of resources from the present organisations
Society and agriculture The unified drought index and the public list of restrictions make the enforcement of the rules more equal Public water abstraction data also contain sensitive economic information, so the level of detail of publication has to be calibrated precisely

The main trade-off is tense between scientific independence and decision relevance. A research institute that is obliged to give expert opinions in official procedures may easily come under political pressure — and precisely for that reason the opinion has to be public: publicity is the only protection that requires no separate institutional guarantee. If the opinion is confidential, the pressure takes effect invisibly as well; if it is public, the departure has to be justified.

The second trade-off is in the timing. The lead time of founding an institution is typically one or two years, whereas the present water situation is measured in weeks. The centre is therefore certainly not an answer to today’s crisis — it is worth saying this clearly, because if the announcement appears as part of daily crisis management, then after the first disappointment it loses the credibility of long-term institution building too. The proposal tips to the risk side if the foundation is aligned with the communications arc of the crisis rather than with its own professional timetable.

Part V — Measurability and summary

5.1 What is worth tracking? (proposed KPIs)

Four performance indicators (KPIs, Key Performance Indicators) are worth tracking in the centre’s first two years:

  1. Completeness of data publication: how many of the data sets listed in the annex of the founding legislation actually appear daily, in machine-readable form — proposed target: 100% by the end of the first year of operation.
  2. Forecast hit rate: the subsequent accuracy of the centre’s water level forecasts over 3, 7 and 14-day horizons, in publicly disclosed form — proposed target: that the measurement itself should start, then an annually improving path.
  3. Expert opinion coverage: in what proportion of above-threshold water abstraction permit applications a public expert opinion was prepared — proposed target: 100% from the second year.
  4. Duplication balance: how many tasks remain that, according to the task map, two or more organisations perform in parallel — proposed target: declining, and the remainder in every case with a named justification.

5.2 Summary

MIAK’s key message in a single sentence: the new water science centre will be durable if the founding decision writes obligations into it, not aims. Concretely, it asks that the founding legislation contain an itemised data publication annex, that for above-threshold water abstraction permits the centre’s expert opinion be compulsory and public, and that within 90 days a public task map be prepared covering the existing players. The judgement of the 2012 closure is secondary in this question: whatever we think of that decision, the present situation also stems from the fact that an institution’s value never became measurable, public performance — and this error can be repeated with the present foundation.

Two MIAK foundational values are in play in this matter. Transparency, because water resources are today the country’s scarcest common resource, and every debate about the distribution of scarcity is settled immediately on the data: as long as the water abstraction permits and the restriction levels are not in one place, publicly, affairs of the stadium watering type remain undecidable, and every player states their own version. And openness, because the natural allies of a state research institute are the universities, the municipalities and the farmers — the mandate must therefore make room not only for the preparation of government decisions but also for external access; knowledge to which only the commissioning party has access does not outlive a single government term.

Part VI — Justifications and further sources

6.1 The press framing by spectrum

The economic band was the only one that made institutional news out of the day. Portfolio brought the announcement in a stand-alone article, with the history of the closed research institute, a description of the fragmentation of its tasks and the elements of the planned mandate — satellite observation, artificial intelligence, sensor networks, digital catchment models. The same paper also published the news of the large-city water restrictions and of the situation in the capital, thus keeping the daily operational picture and the long-term institutional question within one band.

The conservative band concentrated on the data and on the possibility of a turnaround. Mandiner brought both the detailed water directorate communication on the record lows on the Drava and the Danube and the fact that the forecasts for the coming days had become somewhat more optimistic — that is, it carried the measurement side, without governmental evaluation. On this day this framing was the closest to the data.

The public affairs and the liberal-left bands highlighted the catchment-level connection and the inequality of enforcement. 24.hu brought the statement of the portfolio head responsible for the living environment that the Danube is not running out in Hungary — historic lows were measured on the German, Austrian and Slovak stretches too — and the water service’s itemised answer to claims circulating online also belongs here, according to which the primary cause of the record low water level is the precipitation deficit forming on the catchment, not the lasting water retention of the barrages. HVG carried the enforcement side: the turf watering going on in stadiums despite the ban, and the working from home and fountain shutdown in Siófok. 444.hu and Népszava brought the regional picture and the inequality of access, the latter on the cistern water use of a Budapest district. From the comparison of the bands one common gap is visible: not a single paper posed the question of what has to be written into the founding decision — the announcement appeared everywhere as news, not as a regulatory task.

6.2 Facts and data

Data Value Source
Foundation of the closed water management research institute 1952 Portfolio, 3 August 2026
Withdrawal of funding and the closure 2012 Portfolio, 3 August 2026
Launch of the fixed-term national laboratory 2022 Portfolio, 3 August 2026
Danube water level at Dunaszekcső relative to the 2018 record −52 cm South Transdanubian Water Directorate, Mandiner, 3 August 2026
Danube water level at Mohács relative to the 2018 record −57 cm South Transdanubian Water Directorate, Mandiner, 3 August 2026
Flow of the Danube on the Mohács stretch 750 m³/s South Transdanubian Water Directorate, Mandiner, 3 August 2026
Flow of the Drava approx. 180–190 m³/s South Transdanubian Water Directorate, Mandiner, 3 August 2026
Danube water level Paks / Esztergom / Baja −137 / −38 / −24 cm National Directorate of Water Management, 24.hu, 3 August 2026
Water reserve deficit of Lake Balaton 138 million m³ National Directorate of Water Management, 24.hu, 3 August 2026
Water reserve deficit of Lake Velence 23 million m³ National Directorate of Water Management, 24.hu, 3 August 2026
Below-record water level on the upper stretch Hofkirchen −22 cm, Pfelling −6 cm, Kienstock −1 cm relative to the earlier negative record 24.hu, 3 August 2026
Precipitation expected for Thursday on the Inn, Traun and Enns catchments 10–20 mm National Directorate of Water Management, HVG, 3 August 2026
Kaposvár’s network development HUF 70 million, 2 new wells, +1,000–1,300 m³/day on top of the 14 thousand m³ capacity Portfolio, 3 August 2026
Kaposvár’s leak forecasting system 450 sensors Portfolio, 3 August 2026

A single connection deserves highlighting. Today’s data arrived from three different institutions — from the national water service, from a regional water directorate and from the portfolio head’s social media page — in three different forms, and not for the same day. This is not negligence but the direct consequence of the post-2012 institutional structure: the data exist, but there is no single interface where they come together. The proposed data publication annex eliminates precisely this gap — it does not ask for new measurement, but for the already existing measurements to be available in one place, in one format, with one timestamp.

6.3 Policy dimensions

  • Environment and climate (programme points) — the public interface of water quality and water resource monitoring, and integrated climate-economic modelling as an institutional product (programme point ID: K3, K1, K9);
  • Public administration and e-government (programme points) — the measurable replacement of parallel organisational tasks and the delimitation of competences (programme point ID: KI3);
  • Agriculture (programme points) — the data on irrigation water abstraction and the data needs of precision farming (programme point ID: MG2, MG1);
  • Territorial inequality and rural policy (programme points) — the territorial pattern of drinking water outages and the rural public service minimum (programme point ID: TE4);
  • Digitalisation and AI regulation (programme points) — the requirement of machine-readable publication and of a metadata standard (programme point ID: D2, D18).

6.4 Literature in detail

6.4.1 Daniel Kahneman: Thinking, Fast and Slow

In one of the volume’s most personal chapters Kahneman derives two of his basic concepts from the failure of his own textbook-writing project. According to the story, the authors estimated two years for the completion of the work, while the actual lead time of similar undertakings was many times that. From this he drew three lessons: he recognised the difference between the inside and the outside view of forecasting, the concept of the planning fallacy for estimates that are “unrealistically close to the best-case scenario”, and what he called irrational perseverance — that even on the day of the recognition the group preferred to go on rather than give up the plan.

In the closing part of the book an expressly institutional conclusion is added to this, and from the point of view of today’s subject that is the decisive one. According to the author, organisations are better than individuals at avoiding errors, because by their nature they think more slowly and have the power to prescribe regulated procedures — they are able to introduce and require checklists, or to apply more thorough methods such as reference class forecasting and premortem analysis. A water science centre is exactly such an organisation: its task is not to employ cleverer experts than the university faculties, but to make compulsory those procedures that no single expert can keep to on their own — among them the regular, public back-measurement of its own earlier forecasts. The second half of proposal 3.2 is therefore not an administrative burden but the institutional realisation of Kahneman’s methodology.

📖 Source: Daniel Kahneman: Thinking, Fast and Slow

6.4.2 Daron Acemoglu – James A. Robinson: Why Nations Fail

The authors’ central distinction stands between inclusive and extractive institutions. By their definition an economic institutional system is inclusive if it allows and encourages people to take part, in as large a proportion as possible, in economic activities in which they make the best use of their abilities — and for this, alongside the security of private property and the impartiality of the legal system, they list as an explicit element the provision of public services, which “creates a level playing field”. According to the volume’s argument, inclusive institutions increase economic activity and productivity, while extractive ones place obstacles in the way of growth; the decisive consideration is always whether the rule and the access apply equally to every member of society.

This frame is directly applicable to Hungarian water management at two points. One is data access: if the data on water abstraction permits and restriction levels are available only at the acting organisations, then the conditions of access to the scarce resource are not the same — whoever gets hold of the information is able to assert their interest. The other is enforcement: the present stadium watering affair is a policy question and not a tabloid item because if the same ban does not take effect on certain players, then the restriction is not a public service rule but a bargaining position. MIAK derives from this not a tightening but a measurement: the level playing field can be checked if the water abstraction data appear for everyone in the same register. In MIAK’s proposal this is the new centre’s first and most easily fulfilled task.

📖 Source: Daron Acemoglu – James A. Robinson: Why Nations Fail

6.5 International comparison

The Dutch example is the closest template: the country’s water and subsurface research is brought together by an independent applied research institute operating with a public-benefit mandate, Deltares, which makes its models and professional materials publicly accessible to a significant extent. The lesson of the model is not the organisational form but the mandate: the institute is set up for public-interest research, and alongside decision preparation its task is also the public sharing of knowledge — precisely the duality that proposal 3.1 asks for.

The EU level already supplies today what the domestic system does not: the European Drought Observatory of the Copernicus programme publishes drought indices with a unified methodology, publicly and in comparable form, while for the Danube catchment the International Commission for the Protection of the Danube River (ICPDR) maintains joint data collection. The Hungarian institution therefore does not make good missing data, but missing domestic resolution and domestic responsibility — an EU drought index cannot say which restriction level is in force in which Hungarian municipality, and how large a permitted water abstraction is relative to the local reserve.

Environment and climate

  • K3 — Pollution monitoring
  • K1 — Measurable climate targets
  • K9 — A domestic integrated climate-economic model

Public administration and e-government

  • KI3 — Bureaucracy reduction, measurably

Agriculture

  • MG2 — Agricultural data platform
  • MG1 — Precision agriculture programme

Territorial inequality and rural policy

  • TE4 — Rural public service minimum

Digitalisation and AI regulation

  • D2 — Open data programme
  • D18 — Data governance and open data intermediaries

Proposed new programme point: An institution-founding data publication minimum — for the Public administration and e-government area: the founding legislation of every new state research or analytical institution should contain an itemised annex on the data sets compulsorily published in machine-readable form and on the frequency of publication, together with a public task map of its relationship to the organisations already operating.

6.7 List of sources

Press sources (MIAK press monitor, 4 August 2026 — topic 2):

Knowledge-base references (books):

  • 📖 Daniel Kahneman: Thinking, Fast and Slow
  • 📖 Daron Acemoglu – James A. Robinson: Why Nations Fail

Note: the local file path of the books does NOT appear in the visible text of the blog — only the author and the title. The file path is an internal matter of the generation process, not the reader’s.

MIAK internal materials:

  • MIAK policy area: Environment and climate (programme points; programme point ID: K3, K1, K9)
  • MIAK policy area: Public administration and e-government (programme points; programme point ID: KI3)
  • MIAK policy area: Agriculture (programme points; programme point ID: MG2, MG1)
  • MIAK policy area: Territorial inequality and rural policy (programme points; programme point ID: TE4)
  • MIAK press monitor, 4 August 2026 — topic 2, score: 92/100

Additional public data sources:

  • Copernicus European Drought Observatory — unified European drought indices
  • ICPDR — international data collection for the Danube catchment
  • National Directorate of Water Management — water level and flow data
  • KSH (Hungarian Central Statistical Office) — water management and water use statistics
  • Eurostat — water abstraction data broken down by sector

Generation metadata