Part I — Situation overview

The last generating unit but one of the Paks nuclear power plant was shut down at half past one in the early hours of 2 August 2026, and with this the complete shutdown of the plant began. According to the plant’s own statement, after this only unit 2 was operating, at fifty per cent output, and even that only for as long as the Danube’s water level allows the safe operation of the cooling pumps. The plant supplied about forty per cent of Hungary’s electricity generation; in the facility’s 44-year history a complete shutdown has never been necessary. The lead-up is a matter of days: on Thursday unit 3 shut down, by Friday capacity had fallen below the nominal 2,000 megawatts to about 960 megawatts, and the date of the shutdown moved forward day by day: from the time indicated for Tuesday or Wednesday it became Sunday.

In parallel with the plant’s shutdown the legal framework also changed. At 7 p.m. on 1 August 2026 the government decree published in the Hungarian Official Gazette entered into force, replacing the previous order of curtailment. The decree introduces special rules for large consumers above 0.5 megawatts: in a crisis situation they may be required to reduce their consumption, they must comply, and they can have no claim for compensation. If they fail to comply, they may pay a penalty and in the last resort be disconnected from the grid. The decree distinguishes two crisis levels, both tied to measurable thresholds: one case of level I is when the remaining capacity of the electricity system falls to between 750 and 250 megawatts, and level II when it falls below 250 megawatts. In addition, electricity market transactions may be suspended in a crisis situation, the feed-in of household solar systems is temporarily permitted more widely, and from Monday rail freight traffic will stop between 5 p.m. and 10 p.m.

MIAK’s reading: with yesterday, Hungarian energy crisis management received what had until now been missing — curtailment has acquired a basis in a legal source and a numerical triggering threshold. This is a substantive step forward. At the same time the decree regulates one half of the story: how consumption can be forced downwards. The other half — under what measured conditions production restarts, and how long the curtailment lasts — has remained unregulated. According to the technical management, restarting four shut-down, cooled units may take as much as a week, and cannot even begin until sufficient cooling water is available in the Paks area — not even if rain is already falling in the catchment. The country’s economic damage today is therefore caused not primarily by the lost megawatts but by the fact that the timing of the return cannot be planned by anyone. This is not a technical but a procedural gap, and it can be remedied at exactly the level at which the curtailment decree was born last night.

Part II — Literature foundation

Before turning to MIAK’s proposals it is worth fixing the conceptual frame. János Kornai (economist of Hungarian origin, developer of the theory of the soft budget constraint; Harvard professor between 1986 and 2002), in his 1980 monograph Economics of Shortage, described systematically what happens when the allocation of a resource passes from prices to administrative decision: the hoarding tendency appears — actors build up reserves because of uncertain resupply — and hoarding itself deepens the shortage, closing the process into a self-reinforcing circle. The World Bank’s report World Development Report 2015 adds the operating condition of the voluntary phase: most people and organisations are conditional cooperators, that is, they are willing to make a sacrifice as long as they see that others are making one too — so the success of voluntary economising turns not on the strength of the appeal but on visible feedback about the performance of others. And the International Monetary Fund’s (IMF) World Economic Outlook 2025 reminds us that energy security spending in Europe is no longer a cyclical item but a lasting structural risk burdening the public debt path. The detailed treatment of the literature — author by author, with quotations — can be found in the 6.4 Literature in detail section.

📖 Source: János Kornai: Economics of Shortage (1980); World Bank: World Development Report 2015 — Mind, Society, and Behavior; IMF: World Economic Outlook 2025

Part III — MIAK’s concrete proposal

MIAK proposes three measurable measures, every one of which can be carried out even during the shutdown now under way, and none of which requires new legislation.

3.1 A public restart protocol with measured thresholds (within 7 days)

Yesterday’s decree defines two crisis levels with numerical thresholds — remaining system capacity of 750 and 250 megawatts respectively — so the legislator was capable of describing the entry condition in a measurable way. According to MIAK the same technique must be applied to exit: within 7 days it should be published under what combination of measured parameters unit-by-unit loading restarts. At least three indicators are needed: the water level measured at the Paks gauge, the local flow rate of the Danube and the cooling water temperature — with each one accompanied by the threshold value and the expected duration of holding above the threshold. This should be accompanied by a public, weekly updated start-up timetable showing which unit comes when and how much time elapses from physical start-up to nominal output. This is not a promise and not a commitment to a date: it is a conditional schedule, which is informative even if the condition is not met for weeks — because the company at least knows what to watch. The proposal is linked to the K7 energy market shock resilience programme point, and through the planning treatment of cooling water exposure also to K2 energy transition planning.

3.2 A daily updated public saving dashboard (within 14 days)

Undertakings of 240 megawatts have so far been received for voluntary consumption reduction, and according to government statements performance is being monitored continuously. The result of that monitoring is, however, not public today. MIAK proposes that within 14 days a daily updated publication, downloadable also in machine-readable form, should be launched, showing three data items: the daily peak of system load, the actual saving from voluntary and from mandatory curtailment in megawatts, and the fulfilment rate of the undertakings by consumer category — without company names, but broken down by category. This is not a communication tool but the operating condition of the programme: in the logic of conditional cooperation (see 6.4.2) the making of sacrifices is sustained if it is visible that others are making them too. In the present situation no feedback whatever reaches the population and small businesses about their economising effort — so there is nothing to measure against, and nothing to sustain it. The publication fits the logic of the D9 decision-support system and the A1 public data disclosure.

3.3 The counterpart of excluded compensation: a uniform compensation rule (within 30 days)

The decree states that a large consumer subject to mandatory curtailment can have no claim for compensation. In a crisis situation this is a defensible legislative decision — system operation cannot function with an eye on litigation — but according to MIAK it is one-sided in itself. If the state places the burden of curtailment unilaterally, then it must also write down the rule for sharing that burden: within 30 days a public compensation regime, identical for everyone concerned, should be drawn up, fixing what grid use fee discount, capacity payment or offset on the tax side is due for curtailed consumption, and with what documentation it can be called down. There are two reasons for this. One is legal: the exclusion of compensation and the absence of any offset together render the intervention disproportionate, and proportionality is a basic requirement of restrictive legislation. The other is economic: unpredictable curtailment without compensation encourages companies to build up reserves and to postpone production pre-emptively (see 6.4.1) — that is, it creates precisely the shortfall the curtailment is meant to avoid. The proposal is linked to the G9 strategic industrial policy and the G25 energy price shock preparedness plan programme points.

The three proposals are bound together by a single principle: crisis management is not only about ordering curtailment, but also about those curtailed knowing how long it lasts, what their sacrifice is worth and what they get for it. Yesterday’s decree answered the first question; the other two remained open, and both can be closed without money and without amending legislation.

Part IV — Expected effects and risks

Dimension Expected effect Risk
Economy The conditional restart schedule reduces corporate stockpiling and pre-emptive production postponement; the compensation regime makes the cost of curtailment predictable Compensation means a budgetary burden alongside an additional expenditure of at least 50 billion forints in any case; the published thresholds may prove too narrow after the fact
Society The daily updated saving figure sustains the cooperation of households and small businesses, and protects those economising from the feeling that they are straining in vain Making poorly performing categories public can also be used to set communities against one another, if the breakdown is too fine
Public administration Fixing the exit condition reduces the scope for discretion, and makes it verifiable after the fact how long the curtailment was justified Publishing the thresholds creates political pressure for an early restart; nuclear safety consideration cannot be subordinated to a timetable

The main trade-off is tense between being bound and having room for manoeuvre. A restart threshold promulgated in advance takes away the decision-maker’s flexibility: if the measured value is met, they have to give an explanation of why they are not restarting. This is uncomfortable, but that is precisely its benefit — today it works the other way round: the decision can be postponed without explanation. The proposal tips towards the risk side if the public threshold forces the operator to move faster than nuclear safety consideration allows. A single element excludes this: it must be fixed in the protocol that meeting the threshold is a necessary but not a sufficient condition of start-up, and that the licensing authority’s decision has priority in every case.

The second trade-off is the level of detail of the data publication. The daily saving expressed in megawatts is in itself harmless and useful; a breakdown by consumer category, on the other hand, may already make large players identifiable in a given region. MIAK therefore proposes publication at category level, not at player level — the aim is to make the system’s performance visible, not to shame.

Part V — Measurability and summary

5.1 What is worth tracking? (proposed performance indicators)

According to MIAK the following performance indicators (KPIs, in English Key Performance Indicator) are worth tracking:

  • The appearance and completeness of the restart protocol: whether the measured thresholds are published within 7 days, and whether all three parameters (water level, flow rate, cooling water temperature) figure in them — the yardstick is the presence of a numerical value, not a general reference to a “favourable water situation”.
  • The continuity of the saving data publication: on how many days the daily saving figure appeared in the first 30 days after introduction — the proposed target being complete daily publication.
  • The fulfilment rate of voluntary undertakings: how much of the announced 240 megawatts of undertakings was actually realised, measured as a peak-period average — this shows whether the voluntary phase is real capacity or merely a declaration of intent.
  • The actual shutdown time compared with the estimate: the number of days elapsed from the beginning of the restart to reaching nominal output, as against the one-week horizon indicated by the technical management — proposed publication after the fact, 30 days after the close of the shutdown.

5.2 Summary

The key message: MIAK asks the government, after creating the legal framework of curtailment, to close the other half as well — to promulgate the measured conditions for switching back on within 7 days, to launch the daily saving data publication within 14 days, and to publish the uniform rule of compensation alongside the exclusion of damages within 30 days. None of them is an energy policy statement about whether nuclear energy is a good direction; all three are procedural questions, and all three can be carried out during the crisis.

Two MIAK foundational values are in play in this matter. Accountability, because the decree names the burden — mandatory reduction, penalty, disconnection, excluded compensation — but not the temporal limit of that burden: where the end of curtailment is not tied to a condition, it cannot be established after the fact whether on a given day the curtailment was still necessary or merely convenient. And data-drivenness, because the whole of the voluntary phase rests on a single assumption: that people and companies will economise. This assumption is tenable only if the result of the economising is measured and published — without this the appeal is not public policy but a request for trust. MIAK asks the same yardstick of every government: a crisis situation is not a dispensation from measurement but the moment when the absence of measurement costs the most.

Part VI — Justifications and further sources

6.1 The press framing by spectrum

The liberal-left band made uncertainty the centre of its framing. 444.hu’s analysis entitled The catastrophe is not the shutdown of Paks but that we do not know when it can be switched back on, built on expert statements, expressly named the unknown time horizon of the return as the main stake, and with the words of Attila Aszódi, former government commissioner for Paks, emphasised that not a single model had forecast the present flow rate. Telex, by contrast, remained in an event-following register: it gave factual news of the shutdown of the last generating unit but one and of the entry into force of the decree, but it did set out the content of the legislation in detail — including the threshold figures of the crisis levels. The band thus gave both the expert reading of the risk and the precise content of the legislation.

The economic and public-affairs band concentrated on the operational decisions and the capacity balance. Portfolio listed the elements of the prime ministerial announcement in the greatest detail — the 240-megawatt total of voluntary undertakings, the order of the curtailment chain, the temporary widening of solar feed-in, the evening stopping of rail freight traffic and the ordering of home working in public administration — and carried in a separate item the ahead-of-schedule return of unit G3 of the Dunamenti power plant. 24.hu placed the emphasis on the text of the decree and on explaining why household solar systems do not work during a power cut. This band therefore gave the most implementation detail, but did not substantively raise the question of the restart.

The conservative band chose the frame of unprecedentedness and of an event of historic scale: Magyar Nemzet under the title Unprecedented events have begun in the history of the Paks nuclear power plant, Mandiner reporting by following the steps of the gradual output reduction. One thing was missing from every band of the spectrum: that element of the decree’s text which excludes the compensation claims of curtailed large consumers. Telex reported it as a fact, but it received neither critical nor defensive comment anywhere — even though this is the decree’s provision with the strongest effect on property.

6.2 Facts and data

Data Value Source
Shutdown of the last generating unit but one 2 August 2026, 1.30 a.m. statement of MVM Paks Nuclear Power Plant, Telex, 1 August 2026
Entry into force of the government decree 1 August 2026, 7 p.m. Hungarian Official Gazette, Telex, 1 August 2026
The large-consumer threshold falling under the decree above 0.5 MW Telex, 1 August 2026
Crisis level I — remaining system capacity band between 750 and 250 MW Telex, 1 August 2026
Crisis level II — remaining system capacity below 250 MW, above 0 MW Telex, 1 August 2026
Paks’s share of domestic electricity generation approx. 40% Telex, 1 August 2026
The plant’s nominal output 2000 MW 444.hu, 1 August 2026
Output on the day of the announcement 480 MW Portfolio, 1 August 2026
Voluntary undertakings received 240 MW Portfolio, 1 August 2026
Return of unit G3 of the Dunamenti power plant 1 August 2026, 4.45 p.m., approx. 360–400 MW Portfolio, 1 August 2026
Estimated time needed for the complete restart as much as a week statement of MVM’s deputy CEO for technical affairs, MTI/Portfolio, 31 July 2026
Physical start-up of one unit 1–2 days, then a further 1–2 days to nominal output MTI/Portfolio, 31 July 2026
Danube flow rate (June 2026 → end of July) 950 → 650 m³/s statement of Attila Aszódi, 444.hu, 1 August 2026
The water intake required for cooling approx. 100 m³/s statement of Attila Aszódi, 444.hu, 1 August 2026
The indicated water level threshold for complete shutdown −133 cm Telex, 1 August 2026
Forecast water level below −144 cm Telex, 1 August 2026
Daytime and night-time national electricity consumption 7000–7500 MW and 3500–4000 MW respectively 444.hu, 1 August 2026
The estimated budgetary effect of the shutdown at least HUF 50 bn (conservative estimate) Telex, 1 August 2026

Among the data a single connection deserves separate highlighting. The restart is difficult because with four cooled units there is no operating unit from which auxiliary steam could be transferred for heating up — the plant therefore heats the system with external electricity. That is, switching back on itself requires electricity, precisely in the period when the system has the least reserve. This interaction in itself justifies the start-up timetable being not improvisation but a schedule published in advance.

6.3 Policy dimensions

  • Environment and climate (programme points) — energy market shock resilience and the planning of the energy transition; the treatment of cooling water exposure as a recurring planning condition (programme point ID: K7, K2);
  • Economy (programme points) — the industrial policy and budgetary handling of curtailment, the obligation of after-the-fact impact assessment (programme point ID: G9, G25, G20);
  • Public administration and e-government (programme points) — the audit of the organisational routines of crisis decision-making: who orders, who implements, who is answerable for the exit decision (programme point ID: KI11);
  • Digitalisation and AI regulation (programme points) — the machine-readable publication of load and saving data (programme point ID: D9);
  • Transparency and anti-corruption policy (programme points) — the public traceability of compensation payments (programme point ID: A1).

6.4 Literature in detail

6.4.1 János Kornai: Economics of Shortage

In his analysis of the shortage economy Kornai devotes a separate chapter to what happens to the purchasing actor when resupply becomes uncertain. According to his description, amid the uncertainties of production and distribution the hoarding tendency appears: the firm accumulates stocks because it does not trust that the necessary input will be available at the required moment. The process is self-reinforcing — in Kornai’s formulation, the stronger the shortage of some input, the more it is hoarded, the more uncertain its resupply becomes, and the stronger the quantity drive grows. The other reaction examined is forced substitution: the buyer does not procure what they would need but what they can get, and this has a price in productivity. Kornai emphasises that central materials allocation as a mechanism does not eliminate these behaviours, it merely gives them a frame — it restricts the actors’ freedom of decision, but does not erase their responses to uncertainty.

Translated to today’s Hungarian situation, this means that introducing mandatory consumption reduction solves nothing in itself if the rule does not make the duration of the curtailment predictable. A company that does not know when Paks restarts, and does not know at what threshold and for how long it will be switched off, will behave in exactly the way Kornai describes: it will postpone energy-intensive shifts, build up reserves, and thereby deepen the shortage itself. The restart protocol is therefore not a question of information but an instrument directly affecting the magnitude of the economic damage.

📖 Source: János Kornai: Economics of Shortage (1980)

6.4.2 World Bank: World Development Report 2015

The report’s second guiding principle — “thinking socially” — rests on the fact that people’s decisions are shaped not only by material incentives but also by social expectations, reciprocity and norms. Citing the results of a public goods experiment covering eight countries, the volume finds that although the proportion of conditional cooperators and free riders differs from country to country, in every society examined conditional cooperators are in the majority — that is, the typical behaviour is that someone cooperates as long as they see that others are cooperating too. From this the report draws the policy conclusion that the design of institutions and the arrangement of incentives are in themselves capable of suppressing or of eliciting the willingness to cooperate.

In the case of the Hungarian energy-saving appeal this has a direct practical consequence. Today households and small businesses economise without feedback: they do not know how much the sacrifice in the evening peak brought in total, nor whether the large consumers are meeting their undertakings. In the logic of conditional cooperation this is the worst arrangement — not because people are selfish, but because they have nothing to which to adjust their own behaviour. Daily publication of saving data is therefore not an informational luxury but the condition of the voluntary phase’s effectiveness.

📖 Source: World Bank: World Development Report 2015 — Mind, Society, and Behavior

6.4.3 IMF: World Economic Outlook 2025

The IMF’s global outlook names separately, among debt path risks, that expenditure pressure arising from ageing, defence spending and energy security — especially in Europe — further worsens medium-term sustainability. The report underpins the transition from fossil energy carriers to renewables not only with climate policy but also with energy security and balance of payments arguments: reducing import dependence tempers the macroeconomic exposure arising from the volatility of fossil markets.

The Hungarian case shows a rare variant of this argument. The present shortfall arises not from import dependence but from the climate vulnerability of a domestic, low-emission generating unit — yet the consequence is the same: the country has to make up the shortfall with expensive imports, and the budget bears an additional burden of at least 50 billion forints. Within the IMF’s frame this means that handling cooling water exposure (deeper water intake works, alternative cooling solutions, reserve capacity) is not environmental expenditure but investment reducing public debt risk — and it must therefore also figure under that heading in the next budget planning round.

📖 Source: IMF: World Economic Outlook 2025

6.5 International comparison

The most direct comparison arises with Romania, where the same cooling water problem appeared in these same days: one unit of the Cernavodă nuclear power plant shut down, and the government prepared at an extraordinary sitting the decision that would, in the event of a state of emergency, allow the Danube’s flow to be diverted at the branching of the Bala and the Old Danube. The comparison yields two lessons. One: according to the Romanian state secretary’s statement the legal possibility of disconnecting industrial consumers already existed there earlier — that is, the Hungarian decree is not a pioneering solution but the belated supply of an instrument customary in the region. The other: the Romanian government is preparing a physical water management intervention to secure the cooling of its own power plant, which directly affects the flow rate of the downstream Hungarian section as well — energy security decisions in this region therefore have cross-border consequences, without a coordinated framework.

In wider European practice, demand response programmes differ from the Hungarian voluntary appeal precisely in that they operate on a contractual basis, with a call-on threshold fixed in advance and with compensation: the consumer keeps their flexibility available not out of goodwill but for a fee. This model gives the template for proposal 3.3 — it asks not for a new institution, but for the domestic regulatory expression of an already established European market instrument.

Environment and climate

  • K7 — Energy market shock resilience
  • K2 — Energy transition plan

Economy

  • G25 — Energy price shock preparedness plan
  • G9 — Strategic industrial policy
  • G20 — Economic policy impact assessment system

Public administration and e-government

  • KI11 — Organisational behaviour audit

Digitalisation and AI regulation

  • D9 — Data-driven decision-support system

Transparency and anti-corruption policy

  • A1 — Public money dashboard

Proposed new programme point: A restoration protocol for critical infrastructure outages — for the Environment and climate area: a set of restart conditions, published in advance and tied to measured thresholds, for the outage of every strategic generating unit.

6.7 List of sources

Press sources (MIAK press monitor, 2 August 2026 — topic 1):

Knowledge-base references (books):

  • 📖 János Kornai: Economics of Shortage (1980)
  • 📖 World Bank: World Development Report 2015 — Mind, Society, and Behavior
  • 📖 IMF: World Economic Outlook 2025

Note: the local file path of the book 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: K7, K2)
  • MIAK policy area: Economy (programme points; programme point ID: G25, G9, G20)
  • MIAK policy area: Public administration and e-government (programme points; programme point ID: KI11)
  • MIAK press monitor, 2 August 2026 — topic 1, score: 94/100

Additional public data sources:

  • Eurostat — inland freight transport performance time series
  • ENTSO-E Transparency Platform — cross-border capacity and system load
  • National Directorate for Water Management — water level and flow rate time series

Generation metadata