Policy does not change where people live, not by 2040. About 54 % of residents live in Riga and its surroundings with or without the packages on this page. What policy changes is how far the services are for those who stay. If nothing changes, access to schools, clinics and council offices falls by 7 % by 2040, because institutions close as people leave.
Keeping the network open holds access slightly above that path. Managed shrinkage pushes it 2.6 % below. Full municipal equalisation gives Latgale 1.5 % better access at no cost to the state budget. The decision on the network is a decision about distance and its price, not about the number of people.
Why it matters
Between 2005 and 2025 the share of Latvia's residents living in Riga and its surroundings rose from 48.6 % to 53.4 %, and Latgale's share fell from 15.9 % to 12.7 %. That means not only that Latgale lost people abroad, but that it lost people to Riga: the young leave to study and work, and rarely come back.
Over the next fifteen years this picture hardly changes, and that is the first thing to know. The distribution of people across regions has inertia: people move rarely, and those who move are mostly young, and there are already few young people in Latgale and Kurzeme. On the base path the share in Riga and its surroundings is 53.5 % in 2040. The uncertainty band in the chart runs from 51.5 to 57.0 %: that is how far the path moves if the two most important assumptions about how strongly the distance to services affects moving are varied within their plausible range. The band is wider than on other pages, and that is honest: how strongly people respond to a closed school has not been measured in Latvia.
Why it matters even if the number of people does not change. Because services shrink faster than people. In the model there are about 2,300 schools, clinics and council offices in the country in 2025, and on the base path 2,000 in 2040, 14 % fewer, while the population falls by about an eighth. For those who stay, the way to school and to the doctor gets longer, and that is the question policy can answer.
Share of people living in Riga and its surroundings
Residents of Riga and the Pieriga region as a share of all residents.
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 to a tenth of a point: the model gives 53.3 % for the start of 2025, the record was 53.4 %. 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
The story of the regions has one inertia, one feedback loop and one assumption about its strength.
Inertia: people move rarely. The distribution of people across regions has inertia: how many people live in Latgale in 2040 is mostly set by how many live there now, how many of them are young and how many children are born. Fifteen years is a short time to change that, and no policy on this page moves people.
The loop: people and services. Fewer people means fewer pupils and patients, so schools and offices close. A longer way to school and to the doctor is one reason to leave, and then there are fewer people still. This is a reinforcing loop: shrinkage creates shrinkage. In the model it works, but not equally strongly in both directions: the number of people affects the number of institutions strongly and quickly, the number of institutions affects people's choices weakly and slowly. So it amplifies a shrinkage that is already happening but does not create it, which is why keeping a school open keeps the services but not the people.
The people-and-services loop. Plus means "the more of one, the more of the other". All pluses in a row means the loop as a whole is reinforcing: shrinkage creates shrinkage. The last arrow, from services close by to staying, is the weakest and the least measured.
The assumption: how strongly people respond to services. The last arrow in the diagram is the one the uncertainty band rests on. If people respond strongly to a closed school, the share in Riga and its surroundings can reach 57 % in 2040; if weakly, it stays at 51 %. In Latvia this has not been measured, and so this page speaks cautiously about the number of people and confidently about services.
Over the next fifteen years inertia dominates. The loop amplifies what is already happening, and policy can act on its middle link, schools and offices, far more reliably than on its last link, people's choices.
What can be done
Four policies that the government, parliament and municipalities 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, and 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, and a package that saves lets the rule cut less.
Protecting the network
What it is
schools, clinics and council offices are not closed in step with the population; state grants to municipalities a tenth larger, and small schools kept open longer.
What it delivers
access to services in 2040 0.4 % better than on the base path, in Latgale 0.4 %. The distribution of people unchanged.
What it costs
2.4 points of GDP over 2028–2045; debt in 2040 +0.3 points. The cheaper of the two that cost money.
How fast
visible in services from 2043, because on the base path closures are gradual and the difference builds slowly.
Managed shrinkage
What it is
the opposite route: institutions close twice as readily as the population falls, municipalities are merged into 25 from 2029, the minimum size of a school is raised by a quarter, grants are cut by a tenth.
What it delivers
a saving for the budget. Access to services in 2040 2.6 % worse, in Latgale 2.6 %. The distribution of people unchanged here too.
What it costs
nothing; it saves 1.3 points of GDP over 2028–2045, and debt in 2040 is 0.2 points lower. The only one that brings the budget money.
How fast
visible in services from 2033.
Full equalisation
What it is
the municipal equalisation fund, that is, the fund through which richer municipalities pass part of their revenue to poorer ones, evens out financial capacity in full instead of in part.
What it delivers
access to services in Latgale 1.5 % better than on the base path in 2040, nationally unchanged, because the money moves from one municipality to another. Abolishing the fund instead would worsen Latgale's access by 3.0 %; see the table of all packages.
What it costs
nothing to the state budget: the money moves between municipalities. Riga and its surroundings pay.
How fast
visible in Latgale from 2031.
Regional housing programme
What it is
a state guarantee for mortgages outside Riga, construction support a quarter larger, and 2,000 social rental flats a year in the regions.
What it delivers
the share of households that spend more than 40 % of their income on housing 0.3 percentage points lower in 2040. The distribution of people unchanged. A side effect: more new homes mean lower prices, and private building outside Riga pays 4 % less well.
What it costs
5.5 points of GDP over 2028–2045, the most of the four; debt in 2040 +1.1 points.
How fast
visible in housing costs from 2035, because the flats have to be built first.
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 revival: protecting the network, full equalisation, better roads and the housing programme together. Managed shrinkage is the opposite route, so it is not part of revival. Rail Baltica is not on this page: it is on the base path as a decision already taken, and what it costs the budget is on the public finances page.
Managed shrinkage. institutions close faster, municipalities merge, grants smaller.
Network protected. institutions not closed in step with the population, grants a tenth larger.
Full equalisation. the equalisation fund evens out municipal finances in full.
Housing programme. a mortgage guarantee, construction support and 2,000 rental flats a year in the regions.
Revival. the network protected, full equalisation, roads and the housing programme together.
People. Here the answer is short: nothing changes. The share in Riga and its surroundings in 2040 is 53.5 % on the base path, 53.3 % with revival and 53.5 % with the housing programme. Protecting the network and managed shrinkage, which are opposite routes, both differ from the base path by less than a tenth of a point.
Share of people living in Riga and its surroundings
Residents of Riga and the Pieriga region as a share of all residents.
Latgale shows the same. Revival raises Latgale's share in 2040 by 0.1 percentage points, the housing programme by 0.0, full equalisation by 0.05. All three are small against the 3.2 percentage points Latgale lost between 2005 and 2025.
Services. Here policy works. On the base path access to services falls by 7 % by 2040. Protecting the network holds access in 2040 0.4 % higher than on the base path, managed shrinkage pushes it 2.6 % lower. These are the only two policies that change access across the whole country, and they run in opposite directions: the question is whether services are kept open for money or closed for a saving.
Access to local services
How close people live to schools, clinics and council offices, relative to today.
Base path
Managed shrinkage
Network protected
Full equalisation
Housing programme
Revival
No comparable recorded series exists for this indicator; only the model's path is shown.
In Latgale the picture is the same but the differences are larger, because there are already few institutions per person there and every closure is felt. Full equalisation on its own improves Latgale's access by 1.5 % in 2040, more than protecting the network does (0.4 %), and it costs the state budget nothing: the money moves between municipalities. Abolishing the fund worsens Latgale's access by 3.0 %. Revival holds Latgale's access 2.0 % better than on the base path, because it contains both.
Access to local services in Latgale
How close people in Latgale live to schools, clinics and council offices, relative to today.
Base path
Managed shrinkage
Network protected
Full equalisation
Housing programme
Revival
No comparable recorded series exists for this indicator; only the model's path is shown.
The benefit. Housing is what the housing programme changes, and it changes it noticeably. On the base path the share of households that spend more than 40 % of their income on housing falls from 7.1 % in 2025 to 7.3 % in 2040, because incomes grow faster than housing costs. The housing programme lowers it by another 0.3 percentage points, revival by 0.3. The gap in municipal revenue per head between regions grows by 1 % on the base path, and no package changes it by more than a tenth of a per cent: these policies do not cure the budget inequality between regions, they compensate for it.
Households overburdened by housing costs
Households that spend more than 40 % of their income on housing.
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 regional packages are small. Before the rule's response, managed shrinkage saves 1.3 points of GDP over 2028–2045, protecting the network costs 2.4, the housing programme 5.5 and revival 8.0; equalisation costs nothing. In 2040 debt, after the rule has absorbed part of it, the differences are −0.2, +0.3, +1.1 and +1.3 points.
Government debt
What the state owes, as a share of what the country produces in a year.
Share of people living in Riga and its surroundings 2040, points
Government debt 2040, points
Equalisation abolished
No equalisation fund: every municipality lives on its own revenue.
+0.2
0.0
Own revenue
Cadastral values revalued to market level, and municipalities receive 80 % of personal income tax.
0.0
+0.5
Roads
Road maintenance a third larger and municipal investment a fifth larger.
−0.1
+0.1
Assessment
Package
Riga and surroundings 2040, points
Services 2040
Services in Latgale 2040
Housing burden 2040, points
Budget cost 2028–2045, points of GDP
Government debt 2040, points
Who gains, who pays
Confidence
Managed shrinkage
+0.01
−2.6 %
−2.6 %
0.0
+1.3
−0.2
the budget gains; those who stay in the periphery pay with a longer way to school and to the doctor
medium
Network protected
−0.01
+0.4 %
+0.4 %
0.0
−2.4
+0.3
those who stay in the periphery gain; the budget pays, a little
medium
Full equalisation
−0.13
0.0 %
+1.5 %
0.0
0.0
0.0
Latgale and the poorer municipalities gain; Riga and its surroundings pay
medium
Housing programme
−0.04
0.0 %
+0.1 %
−0.3
−5.5
+1.1
tenants and buyers in the regions gain; the budget and private builders pay
medium
Revival
−0.26
+0.4 %
+2.0 %
−0.3
−8.0
+1.3
services and housing in the periphery gain; the budget pays most
low
Riga and surroundings, housing burden and debt: difference from the base path in percentage points. Services: difference in per cent. 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. Confidence: how sure the direction and size of the result are, given the assumptions they rest on.
The table says three things. First, the column for Riga and its surroundings is almost empty, and that is the page's main result. The question "how to keep people in the regions" cannot be answered with these means by 2040. Decisions about schools and offices have to be taken knowing that they will not move people one way or the other. Second, the real decision is about services. Protecting the network and managed shrinkage are the same choice seen from its two ends: about 2.4 points of GDP over the period for services 0.4 % closer, or 1.3 points of saving for services 2.6 % further away. Third, the equalisation fund is the one instrument that helps Latgale for free, and it helps Latgale's services more than protecting the network does. It is a decision about who pays within the country, not about the budget.
For a decision-maker this gives a sequence. First the equalisation fund in full, which costs the state nothing. Then the decision on the network, taken as a decision about the distance to services and its price, not about the number of people. And the housing programme where the rental market does not work. Rail Baltica is a transport and public finance decision and is judged on the public finances page.
What to keep in mind
The model holds the distribution of people across regions almost still, and that may be too calm. Between 2005 and 2025 Latgale lost 3.2 percentage points of Latvia's residents; on the base path it loses 0.1 over the next fifteen years. In the model moves between regions follow jobs and services, and both change slowly after 2025. If in reality the outflow from the east continues at its earlier pace, the shrinkage of services in the periphery is faster than shown here, and the value of protecting the network larger.
How strongly people respond to a closed school has not been measured. That is the assumption the wide uncertainty band rests on, and it is the reason this page shows the number of people with a band and services without one. It does not hinder decisions about the network: they act on services, which are reliable, not on people, who are not.
The housing programme lowers the return on private building. More new homes mean lower prices, so private building outside Riga pays 4 % less well in 2040 than on the base path. That is a market working, not an error, but it means the programme partly replaces rather than adds.
The equalisation fund's value to Latgale is proportional to one assumption. How strongly a municipality's financial capacity sets how many institutions it keeps open is an assumption with a wide range in the model. The ranking, that is, full equalisation helps and abolition harms, holds across the whole range; the size can differ several times over.
Rail Baltica is on the base path, not among the packages. The line is a decision already taken and is in the base path with its construction years and its state share. Its gain for the regions, shorter journeys along the corridor, arrives when the line opens in 2035, not during construction. Its effect on where people live is small in the model and is not shown here. Its cost to the budget, and what cancelling it would save, is on the public finances page.
After 2040 the drift continues slowly. The share in Riga and its surroundings is 53.8 % in 2045 on the base path and revival holds it 0.4 points lower; access to services falls by 10 % by 2045. Those years rest on more assumptions than the years before, which is why the page prints 2040.
Several things are not in the model. There is no separate regional quality of education and no travel-time matrix: access is measured by institutions per person and the condition of roads. There is no regional poverty row either. No package changes the at-risk-of-poverty rate or income inequality by a tenth of a point, but that is because the regional layer of the model does not carry the income distribution.
The 2028 start is an assumption. Every package starts in 2028 and reaches full strength in 2030. A later start shifts everything to the right; the result in 2040 is then smaller, but the relations between the policies are the same.