How much energy is lost or wasted before it is used?
Network losses on one definition since 2011, energy intensity on Eurostat's own indicator, and demand per person.
● Refreshed · last verified against source on this date
Scroll the chart sideways to read the full period →
Serbia cut losses from 15.2% to 10.5% across this period. Albania moved the other way, from 14.0% to 17.3%. Two neighbouring systems, one definition, one source, opposite directions. Kosovo has improved only slightly, from 15.1% to 13.7%.
Source: Eurostat — energy balance (nrg_bal_c) (Eurostat open data). The percentage is FlowOps's own calculation — distribution losses divided by transformation output plus imports minus exports, all in GWh from Eurostat — not a ratio Eurostat itself publishes; includes technical loss and, in several of these systems, unbilled consumption. Each series runs to its own last published year, marked on the chart. Between full data refreshes, FlowOps automated monitoring re-confirms this against Eurostat — most recently 1 Sept 2026.
| Country | Year | Losses, percent of output |
|---|---|---|
| Kosovo | 2011 | 15.1 |
| Kosovo | 2012 | 15.2 |
| Kosovo | 2013 | 14.2 |
| Kosovo | 2014 | 13.8 |
| Kosovo | 2015 | 16 |
| Kosovo | 2016 | 21.4 |
| Kosovo | 2017 | 21 |
| Kosovo | 2018 | 18.6 |
| Kosovo | 2019 | 18 |
| Kosovo | 2020 | 14.8 |
| Kosovo | 2021 | 13.7 |
| Kosovo | 2022 | 11.5 |
| Kosovo | 2023 | 10.6 |
| Kosovo | 2024 | 13.7 |
| Albania | 2011 | 14 |
| Albania | 2012 | 15.4 |
| Albania | 2013 | 21 |
| Albania | 2014 | 14.4 |
| Albania | 2015 | 17.1 |
| Albania | 2016 | 25.9 |
| Albania | 2017 | 17.1 |
| Albania | 2018 | 14.6 |
| Albania | 2019 | 13.9 |
| Albania | 2020 | 12 |
| Albania | 2021 | 16.1 |
| Albania | 2022 | 12.1 |
| Albania | 2023 | 13.8 |
| Albania | 2024 | 17.3 |
| North Macedonia | 2011 | 14.7 |
| North Macedonia | 2012 | 14.5 |
| North Macedonia | 2013 | 13.5 |
| North Macedonia | 2014 | 12.8 |
| North Macedonia | 2015 | 12.7 |
| North Macedonia | 2016 | 13.1 |
| North Macedonia | 2017 | 13.2 |
| North Macedonia | 2018 | 13.2 |
| North Macedonia | 2019 | 12.7 |
| North Macedonia | 2020 | 13.4 |
| North Macedonia | 2021 | 13 |
| North Macedonia | 2022 | 13.2 |
| North Macedonia | 2023 | 13.6 |
| North Macedonia | 2024 | 14.1 |
| Montenegro | 2011 | 15.4 |
| Montenegro | 2012 | 17.1 |
| Montenegro | 2013 | 21.1 |
| Montenegro | 2014 | 20.1 |
| Montenegro | 2015 | 16.6 |
| Montenegro | 2016 | 15.7 |
| Montenegro | 2017 | 14.2 |
| Montenegro | 2018 | 13.9 |
| Montenegro | 2019 | 13.4 |
| Montenegro | 2020 | 14.1 |
| Montenegro | 2021 | 14 |
| Montenegro | 2022 | 15 |
| Montenegro | 2023 | 14.3 |
| Montenegro | 2024 | 13.9 |
| Serbia | 2011 | 15.2 |
| Serbia | 2012 | 15.1 |
| Serbia | 2013 | 14.7 |
| Serbia | 2014 | 14.5 |
| Serbia | 2015 | 13.8 |
| Serbia | 2016 | 12.8 |
| Serbia | 2017 | 12.7 |
| Serbia | 2018 | 12.1 |
| Serbia | 2019 | 11.5 |
| Serbia | 2020 | 11.7 |
| Serbia | 2021 | 11.5 |
| Serbia | 2022 | 10.9 |
| Serbia | 2023 | 10.9 |
| Serbia | 2024 | 10.5 |
| Bosnia and Herzegovina | 2014 | 9.3 |
| Bosnia and Herzegovina | 2015 | 9.7 |
| Bosnia and Herzegovina | 2016 | 9.7 |
| Bosnia and Herzegovina | 2017 | 9.2 |
| Bosnia and Herzegovina | 2018 | 9.3 |
| Bosnia and Herzegovina | 2019 | 9.1 |
| Bosnia and Herzegovina | 2020 | 9.6 |
| Bosnia and Herzegovina | 2021 | 9.7 |
| Bosnia and Herzegovina | 2022 | 9.5 |
| Bosnia and Herzegovina | 2023 | 9.6 |
| Bosnia and Herzegovina | 2024 | 9.4 |
The decision behind this figure
Losses respond to metering, billing, enforcement and network investment — Serbia and Albania moved in opposite directions on the same definition. Each percentage point recovered is capacity that needs no connection agreement.
04 · Energy Efficiency, Buildings & InfrastructureThe cheapest megawatt is the one nobody has to generate.
Every transition debate in this region is about what to build. Rather less of it is about what is already being lost — in networks, in buildings, and in the energy an economy burns per unit of output. Those losses are recoverable capacity, usually at a fraction of the cost of new generation, and they do not need a connection agreement.
Three indicators, each with its own publication year, none of them ranked against the others. The reason is directly below the figures.
Network losses
% of electricity generated, lost before the meter
- Kosovo13.7%· 2024
- Albania17.3%· 2024
- North Macedonia14.1%· 2024
- Montenegro13.9%· 2024
- Serbia10.5%· 2024
- Bosnia & Herz.9.4%· 2024
Distribution losses as a share of the electricity that entered the system. Includes technical loss and, in several of these systems, unbilled consumption — which is why this reads as much as a governance measure as an engineering one. Percentage derived from published GWh; see the note below.
Energy intensity
kgoe per EUR 1 000 of GDP · Eurostat indicator
- Kosovo404· 2022
- Albania166· 2024
- North Macedonia267· 2024
- Montenegro253· 2024
- Serbia357· 2024
- Bosnia & Herz.379· 2024
How much energy the whole economy burns per unit of output. This is Eurostat's own published indicator on chain-linked volumes — constant prices — not a FlowOps calculation. The EU-27 reads 100 in 2024. Constant prices matter here: on a nominal-GDP denominator a lower-price economy reads as more energy-intensive partly because of its price level rather than its energy use. Years differ by country and each is printed.
Demand per person
MWh of electricity per person per year
- Kosovo4.43· 2025
- Albania2.88· 2023
- North Macedonia4.68· 2025
- Montenegro5.52· 2025
- Serbia5.68· 2025
- Bosnia & Herz.3.90· 2025
Read alongside the other two, never alone: low per-person demand can mean an efficient system or an under-supplied one, and the difference matters entirely for what an intervention should be.
A second calculation, ours, shown because it is computed differently. Gross available energy (nrg_bal_c, GAE) ÷ GDP at current prices (nama_10_gdp, B1GQ), 1 ktoe = 11.63 GWh, same year both sides. On this basis Kosovo reads 3.07 kWh per euro of GDP in 2024 (2,731.5 ktoe ÷ €10,350.4m), so the arithmetic can be checked by hand. Nominal-GDP denominator. Cross-country comparison on this basis is affected by price level as well as by energy use, which is why the official chain-linked indicator is the one shown.
These three indicators are not interchangeable, and none of them ranks a country on its own. Network losses are as much a billing and enforcement measure as an engineering one. Energy intensity moves with industrial structure, not only with efficiency. Low demand per person can mean an efficient system or an under-supplied one. Rows stay in regional order rather than sorted, because a league table across three unlike measures would imply a single score that does not exist. Each value carries its own year and source.
Sources: Eurostat — energy balance (nrg_bal_c) (Eurostat open data) · Eurostat — energy intensity of the economy (nrg_ind_ei, chain-linked volumes) (Eurostat open data) — Eurostat's published indicator on Eurostat's methodology, not a FlowOps calculation. Unit: kgoe per EUR 1 000 of GDP. Network losses use FlowOps's own supply-basis calculation, not a ratio Eurostat itself publishes — see the losses chart for the formula. · demand per person from Ember and Our World in Data as shown elsewhere on this page. Each value carries its own year. Between full data refreshes, FlowOps automated monitoring re-confirms these against Eurostat — losses checked 1 Sept 2026, intensity checked 1 Sept 2026.
| Country | Network losses, percent | Losses year | Energy intensity, kWh per unit GDP | Intensity year | Demand per person, MWh | Demand year |
|---|---|---|---|---|---|---|
| Kosovo | 13.7 | 2024 | 404.1 | 2022 | 4.43 | 2025 |
| Albania | 17.3 | 2024 | 165.9 | 2024 | 2.88 | 2023 |
| North Macedonia | 14.1 | 2024 | 266.5 | 2024 | 4.68 | 2025 |
| Montenegro | 13.9 | 2024 | 252.8 | 2024 | 5.52 | 2025 |
| Serbia | 10.5 | 2024 | 356.8 | 2024 | 5.68 | 2025 |
| Bosnia and Herzegovina | 9.4 | 2024 | 379.2 | 2024 | 3.9 | 2025 |
The decision behind this figure
Losses, energy intensity and demand per person each point at a different intervention — network investment, industrial structure, or building stock. Which one is worth funding first is a modelling question, not a ranking of these three numbers.
04 · Energy Efficiency, Buildings & InfrastructureWhich measures to fund, in what order, and what each one actually returns per euro and per tonne.