What does cutting emissions cost Latvia, and where is the money best spent?
About 18 min read
The short answer
Emissions from energy fall by themselves, because the wind farms already planned get built. If nothing changes they fall by 10 % between 2025 and 2040. No package on this page cuts them by more than another 2.5 %, because what remains sits in transport and farming. What the transition changes is the price. Wholesale electricity falls to about 47 euro per megawatt-hour by 2040, close to half today's. Heat pumps and more power plants decide who pays and how much.
Latvia's real climate bill is the forest. Counting forests and land, net emissions rise, because the forest now releases carbon instead of absorbing it. The only policy that changes that is less felling: net emissions 8.8 % lower in 2040. Its price to forestry is one the model does not know.
Why it matters
Latvia has two obligations in one: to cut emissions to the European Union's targets, and to buy allowances from other countries if it misses them. Between 2005 and 2025 emissions without forests fell from 11.2 to 9.6 million tonnes, while the energy used rose from 155 to 180 petajoules: the economy grew, and the carbon in each unit of energy fell. That is the pattern of the next fifteen years too.
On the base path emissions without forests are 8.6 million tonnes in 2040. The uncertainty band in the chart runs from 8.4 to 8.8 million: that is how far the path moves if the two most important assumptions about economic growth are varied within their plausible range. The band here is about growth, not about energy: the faster the economy grows, the more is driven and produced.
Why it matters even though emissions fall. Because the European Union's targets for Latvia count forests too, and there the trend runs the other way. The balance of forests and land in 2005 was 14 million tonnes of carbon absorbed. In 2025 it is 3.2 million tonnes released, because the forest is older, felled more intensively, and a considerable part of what is felled goes into heating. On the base path this grows to 4.3 million in 2040. The state buys the difference between target and reality. Buying allowances from other countries costs up to 240 million euro a year around 2031 in the model, in today's money, that is, at 2025 prices. That is under the current targets, which end in 2030. If the targets continue after that in the same form, the bill is many times larger; see "What to keep in mind".
Greenhouse gas emissions
Emissions from energy, transport, industry, farming and waste, excluding forests and land.
Recorded data
Base path
Uncertainty band
The grey band shows how far the base path moves when the two most important assumptions are varied across their plausible ranges.
In the chart the dots are the recorded data and the line is the model's base path. Up to 2025 they match: the model gives 9.6 million tonnes for the start of 2025, the record was 9.6. That is a test: if the model can repeat the past, there is reason to listen to what it says about the future.
What drives it
In energy and climate there is no single loop that decides everything. There are three long-lived systems, one assumption about what has already been decided, and one forest.
Three systems that change slowly. A power plant serves thirty years, a boiler twenty, a building a hundred. So how much Latvia emits in 2040 is largely decided by what is being built now, and every policy has a delay between decision and effect. Electricity changes fastest. The wind and solar farms already under construction bring the wholesale price in the model down from 86 to about 45 euro per megawatt-hour by 2035, and Latvia turns from importer to exporter. The share of renewable energy in all the energy used rises from 43 % to 58 %. Energy imports fall from 36 % to 22 % of the energy used. Heat changes slowly: wood supplies 68 % of heating now and 76 % in 2040 on the base path, because wood is cheap and local. Buildings change slowest of all: about one per cent of the floor area is renovated a year, which is why the energy used stays flat while the economy grows by a fifth.
The assumption: the planned wind farms get built. The base path assumes that the wind farms with permits and contracts are finished on time. That is a judgement about construction, not a forecast. If only half is built, electricity in 2035 is 7 % dearer and emissions higher, and the table of all packages shows that as a separate case.
The forest. The wood that replaced gas in heating is the same wood that turned the forest from absorbing carbon to releasing it. This is the heaviest figure on the page. The balance of forests and land in 2040 is 4.3 million tonnes of carbon released, half of all other emissions together. No heat pump changes that materially. Only less felling changes it, and the model has no economic price for that: forestry and wood processing are a large export industry, but the model does not account for it separately. That is the largest thing missing from this page.
Over the next fifteen years what has already been decided dominates: electricity moves to wind by itself, heat stays with wood, buildings renovate slowly. Policy can make electricity cheaper, speed up heat and buildings, and choose how much of the building bill the state carries. The forest is a separate decision, for which only the climate account is shown here, not the economic one.
What can be done
Four policies that the government and parliament can actually introduce. For each: what it delivers in 2040 compared with the base path, what it costs the public budget, and how fast it acts. The cost to the budget is the package's own effect on the budget balance, added up over 2028–2045 in points of gross domestic product (GDP, that is, everything the country produces in a year). The difference in government debt in 2040 is shown beside it. It is smaller, because under the European Union's fiscal rules a package that spends is mostly paid for by cuts the rule forces elsewhere.
More power plants
What it is
wind and solar capacity built half as fast again as planned, storage added, and a further interconnection with the Nordic market.
What it delivers
the wholesale electricity price in 2040 10 % lower than on the base path, from the first years; emissions without forests 0.9 % lower in 2040 and 1.1 % lower in 2033, while gas is still being displaced. The share of renewables in the energy Latvia uses hardly changes, because the extra power is exported.
What it costs
nothing to speak of: 0.0 points of GDP over 2028–2045, because the plants are private and the budget sees them through taxes; debt in 2040 0.0 points.
How fast
in the price from 2027, as soon as the first plants connect.
Heat pumps
What it is
support for heat pumps, that is, electric heating that takes more heat from the outside air or the ground than the electricity it uses, 2.5 times larger than today, and a faster move away from wood and gas boilers.
What it delivers
emissions without forests 1.2 % lower in 2040, the share of heat from wood 2 percentage points smaller, energy imports 0.4 points lower. The side effect is the point: electricity demand in 2040 is 4 % higher and the wholesale price 6 % higher, which is why heat pumps and more power plants belong together.
What it costs
0.5 points of GDP over 2028–2045; debt in 2040 +0.1 points. Cheap.
How fast
slowly, because boilers are replaced when they wear out: visible in emissions from 2033.
Less felling
What it is
heat pumps as above, and felling 15 % lower, so that less wood goes into heating and more stays in the forest.
What it delivers
emissions from forests and land 24 % lower in 2040, net emissions with forests 9 % lower. The largest climate effect on this page, and the only one that moves the forest.
What it costs
0.5 points of GDP over 2028–2045 in the budget, debt in 2040 +0.1 points, but that is not the whole price: what less felling costs wood processing and exports is not in the model. See "What to keep in mind".
How fast
in the forest at once, from 2029: a tree not felled keeps absorbing carbon the same year.
Renovation grant
What it is
the state pays 70 % of the cost of renovating a building, instead of about half as today.
What it delivers
the share of floor area renovated to a good energy standard 7 percentage points higher in 2040, the energy used 1.6 % lower, energy costs for households 0.3 points of income lower. Emissions without forests only 0.6 % lower: buildings are the most expensive way to cut a tonne.
What it costs
3.5 points of GDP over 2028–2045, the most of the four; debt in 2040 +0.9 points. A side effect: demand for builders raises the price of renovation, and in the first years part of the money buys price, not floor area.
How fast
in renovated area from 2030, in bills from 2034.
What the packages do
Five packages, all introduced from 2028 with a three-year phase-in. Four are one policy each at full strength. The fifth is a balanced policy: more power plants, heat pumps, a mandatory standard for the worst buildings with cheap loans and half the allowance money in grants, and a small carbon dividend. Less felling is not in the balanced policy, because its price is unknown. The question of who pays, the carbon dividend and the bills of the poorest fifth, is in the table of all packages, not on the charts.
More power plants. wind, solar, storage and an interconnection built faster than planned.
Heat pumps. support for heat pumps 2.5 times larger, a faster move away from boilers.
Less felling. heat pumps and felling 15 % lower.
Renovation grant. the state pays 70 % of the cost of renovating a building.
Balanced policy. power plants, heat pumps, a fair building standard and a small dividend.
Emissions. Without forests the differences are small, because the biggest step is already in the base path. A balanced policy cuts emissions in 2040 by 2.5 %, heat pumps by 1.2 %, more power plants by 0.9 % in 2040 and more in the early 2030s, the grant by 0.6 %. What remains after electricity moves to wind is transport and farming, and this page has no policies for those.
Greenhouse gas emissions
Emissions from energy, transport, industry, farming and waste, excluding forests and land.
With forests the picture is different. On the base path net emissions rise, because the forest releases more and more. Less felling cuts them in 2040 by 9 %, a balanced policy by 2.9 %, the others by less than two per cent. This is the figure the European Union judges Latvia by, and it says that the climate question in Latvia is a forest question.
Net emissions including forests and land
Total emissions after what forests and land absorb or release.
Energy demand. The amount of energy Latvia uses is the flattest line on this site. It rose from 155 petajoules in 2005 to 180 in 2025 and stays at about 181 in 2040 on the base path, while the economy grows by a fifth: the energy in each unit of output falls as fast as output grows. Only buildings move it. The balanced policy lowers it by 1.8 % in 2040, the grant by 1.6 %; heat pumps and more power plants leave it where it is, because they change what the energy is, not how much of it is used.
Energy used
All the energy households, transport, industry and services use in a year, after conversion losses.
Inside that total, electricity grows. Demand for electricity is 7.2 terawatt-hours in 2025 and 7.1 in 2040 on the base path; heat pumps raise it by 4 % and less felling, which contains them, by 5 %. That is the demand the new power plants are for, and it is why the two policies belong together.
Electricity demand
All the electricity used in the country in a year.
The transition. The share of renewable energy in all the energy used rises on the base path from 43 % in 2025 to 58 % in 2040, and that is the wind farms already planned. The packages add tenths: heat pumps 0.4 points, the balanced policy 0.6. More power plants add almost nothing to the share, because the extra electricity is exported; what they change is the price.
Renewable energy
The share of all energy used that comes from wind, sun, water and wood.
The wholesale price of electricity, that is, the price on the exchange before network charges and taxes, falls on the base path from 86 euro per megawatt-hour in 2025 to 45 in 2035 and 47 in 2040. More power plants bring it 10 % lower still, heat pumps push it 6 % higher, and the balanced policy, which contains both, ends 6 % below the base path. The table of all packages has two more cases worth reading beside this chart. Offshore wind brings the price down by 38 % but costs the budget 0.7 points of GDP. And if only half the planned wind farms are built, the price is 6 % higher.
Wholesale electricity price
The average price of electricity on the exchange before network charges and taxes.
Buildings and bills. The renovated share is where the grant works: on the base path 45 % of the floor area is at a good standard in 2040, with the grant 52 %, with a balanced policy 53 %. The recorded dots in the chart sit well below the model's line: the model gives 32 % for 2025, the record is 19 %. That is a known level miss, and it is in "What to keep in mind". What every household feels is the bill. Energy costs on average fall on the base path from 8.4 % of income in 2025 to 5.5 % in 2040, because incomes grow and electricity gets cheaper. The balanced policy cuts them by another 0.5 percentage points, the grant by 0.3, heat pumps by 0.07.
The price. On the base path government debt rises from 47 % of GDP in 2025 to 61 % in 2040; why it rises is explained on the public finances page. Against that the energy packages are small. Before the rule's response, more power plants cost the budget nothing, heat pumps 0.5 points of GDP over 2028–2045, less felling 0.5, the grant 3.5 and the balanced policy 8.2; in 2040 debt the rule has absorbed part of it and the differences are 0.0, +0.1, +0.1, +0.9 and +0.9 points. No package materially reduces the purchase of allowances from other countries: that is the forest, and only felling moves the forest.
Government debt
What the state owes, as a share of what the country produces in a year.
Only half of the wind farms already planned are built on time.
+0.2
+0.3
Offshore wind
One gigawatt of offshore wind farms from 2033.
−3.6
+1.7
Interconnectors late
The new electricity links to Poland and Estonia ready four years later.
0.0
0.0
Expensive gas
The gas price almost twice as high as assumed on the base path.
−1.6
−0.2
Cheap renovation loans
Loans for building renovation at 2 % over thirty years.
−0.1
+0.3
Mandatory standard
Buildings in the worst energy classes must be renovated from 2030; owners pay themselves.
−0.2
0.0
Fair standard
The mandatory standard together with cheap loans and half of the emission allowance revenue for renovation grants.
−0.7
+0.9
Carbon dividend
60 % of the revenue from the new emission allowance system is paid out to every resident in an equal sum.
−0.1
+1.0
Allowance money for renovation
60 % of the revenue from the new emission allowance system goes to renovation grants.
−0.6
+0.8
Obligations after 2030
Emission and forest targets keep binding after 2030, and the gap has to be bought from other countries.
−0.1
+2.4
Balanced policy with post-2030 obligations
The balanced policy if the targets keep binding after 2030.
−2.6
+2.8
Less wood for heat with post-2030 obligations
Heat pumps and less felling if the targets keep binding after 2030.
−1.1
+2.1
Assessment
Package
Emissions 2040
With forests 2040
Energy used 2040
Electricity price 2040
Budget cost 2028–2045, points of GDP
Government debt 2040, points
Who gains, who pays
Confidence
More power plants
−0.9 %
−0.6 %
0.0 %
−10 %
0.0
0.0
every electricity buyer gains; private investors pay and earn on exports
medium
Heat pumps
−1.2 %
−1.0 %
+0.2 %
+6 %
−0.5
+0.1
homeowners who replace a boiler and the air in towns gain; the budget pays a little, electricity buyers pay a higher price
medium
Less felling
−1.0 %
−8.8 %
+0.1 %
+7 %
−0.5
+0.1
the climate balance and a state that buys no allowances gain; forestry and wood processing pay, a price not calculated
low
Renovation grant
−0.6 %
−1.2 %
−1.6 %
−1 %
−3.5
+0.9
residents of renovated buildings and builders gain; the budget pays, dear per tonne
medium
Balanced policy
−2.5 %
−2.9 %
−1.8 %
−6 %
−8.2
+0.9
all of the above gain, except the forest; the budget pays most
medium
Emissions, energy used and the electricity price: difference from the base path in per cent, emissions without forests and with forests and land. Budget cost: the package's own effect on the budget balance, added up over 2028–2045, before the EU rule responds; negative costs the budget, positive saves. Debt: difference in percentage points. Confidence: how sure the direction and size of the result are, given the assumptions they rest on.
The table says three things. First, energy emissions fall by themselves, and policy cannot cut them much further. The big decision, wind instead of gas, has already been taken, and this site shows it as the base path. What policy can still buy cheaply is a lower electricity price, and more power plants do that at no cost to the budget. Second, energy demand is not where the transition happens: it stays flat for fifteen years, and only buildings, the most expensive tonne on the page, bring it down. The transition is a change in what the energy is: from imported gas and local wood to local wind. The right pair of policies adds the demand and the supply together, heat pumps and power plants. Third, Latvia's climate bill is the forest, and the only policy that moves it is the only one whose price is unknown. Less felling cuts net emissions more than all the others together. But its effect on wood processing and exports is not in the model, so its confidence is low and it is not in the balanced policy.
For a decision-maker this gives a sequence. First make sure the planned wind farms actually get built, because the base path rests on it, and let more be built, because that is the cheapest thing on this page. Then heat pumps as the cheap and slow way off wood, together with the plants that will power them. The grant for buildings is a decision about bills and construction jobs, not about the climate. And for the forest, a separate calculation that includes its economic side, before taking a decision that looks cheap here.
What to keep in mind
Less felling has no price in the model. Forestry and wood processing are a considerable part of Latvia's exports, but the model does not account for the industry separately, so the package "less felling" barely changes the economy. That is a statement about what the model lacks, not about the policy. The climate side, forest emissions 24 % lower, is reliable; the economic side has not been calculated.
The renovation level in the model is too high. The model gives 32 % of the floor area at a good standard for 2025, the record is 19 %: the model renovates about one per cent of the area a year, in reality it is less. So the grant's effect should also be read as "about this many percentage points above what there would be without the grant", not as a level. The ranking between the grant, the loan and the mandatory standard does not change.
Wind farm construction is an assumption. The base path assumes the planned wind farms are finished on time. The table of all packages has a case in which only half is built: electricity is dearer and emissions higher. The electricity price in the model is an annual average; how expensive the hours without wind will be, the model does not show, and that is worth knowing before the fall in price is taken as certain. The same goes for more power plants: their gain is an average price, and the value of storage and the interconnection is in evening out the hours, which an annual average understates.
Targets after 2030. On the base path the purchase of allowances from other countries falls after 2030, because the current targets end. If the European Union continues the targets in the same form, the forest gap has to be bought every year: the case "obligations after 2030" in the table costs about 620 million euro a year by 2040 in today's money and raises debt in 2040 by 2 points, more than any package. What the targets will be has not been decided, and it is the largest number on this page.
Who pays is in the table, not on the charts. This page is about the economics of the transition. The carbon dividend, that is, paying the revenue from emission allowances out to every resident, and the bills of the poorest fifth are in the table of all packages: the dividend changes no emissions and no prices, it changes who carries the allowance price for heating and motor fuel from 2027. It costs the budget 4.5 points of GDP over 2028–2045 because the money paid out does not stay in the budget.
Expensive gas helps Latvia in the model. The table of all packages has a case with gas almost twice as dear, and in it the economy is slightly larger: expensive gas speeds up wind construction. That is a property of the model, not advice: the first years are dearer for everyone.
The uncertainty band covers two assumptions, not all. They are about economic growth, because that sets how much is driven and produced. Assumptions about wind farm construction, the spread of heat pumps or the cost of renovation are not in the band; the table has separate cases for them.
After 2040 the lines continue in the same direction. Emissions without forests are 8.4 million tonnes in 2045 on the base path and the balanced policy cuts them by 2.7 % then; the forest releases 4.7 million tonnes in 2045, and less felling cuts net emissions by 9.1 %. Those years rest on more assumptions than the years before, which is why the page prints 2040.
The 2028 start is an assumption. Every package starts in 2028 and reaches full strength in 2030; power plants to be connected in 2028 are financed a year earlier. A later start shifts everything to the right; the result in 2040 is then smaller, but the relations between the policies are the same.