Ukraine
What holds the state together, what it lives on, how it handles its own corruption, and which of its numbers will bear weight.
(and roughly $19bn of that was still unsecured when the budget passed)
(the spread is not sloppiness; the last census was in 2001)
(up from 26 in 2014 - the largest gain of any EU candidate - and flat since 2023)
A note on framing. Ukraine appears on eight other pages of this site, and on every one of them it is somebody else's variable: a shock to the wheat market, a source of refugees, a target in the cyberwar section, a precedent in the nuclear section, the trigger for NATO enlargement. This page treats it as the subject. The war is context here rather than content - the fighting itself is covered on the Russia page, and repeating it would add nothing. What follows is about the country as a system: how it is administered, what it lives on, how it handles its own corruption, and who is left in it. Much of the information about Ukraine is produced by people who need it to be either a victim or a fraud, which is exactly why it is worth trying to describe plainly.
How reliable the figures are
Start here, because everything downstream depends on it. Ukraine's last full census was in 2001. Every population figure published since is a projection built on a twenty-five-year-old baseline, and since 2022 the projection has had to absorb occupied territory, mass displacement, military deaths, and a collapse in births, none of which are counted cleanly. The results diverge accordingly. The Institute of Demography of Ukraine's National Academy of Sciences puts the population of government-controlled territory at roughly 29 million, plus or minus a million. Other widely-quoted estimates run to 33 or 35 million depending on whether they count occupied areas, refugees who may return, or registered residents. The World Bank series, the most-cited database in the world, shows about 39 million for 2025.
That last figure is worth dwelling on, because it demonstrates the problem better than any argument could. The World Bank's Ukraine population series does not fall through 2024 and 2025. It rises. Its life-expectancy series for Ukraine reads at an all-time high of about 74.7 years for 2024 - higher than any pre-war year on record - in the middle of the largest war in Europe since 1945. Neither number is fraudulent. Both are modelled outputs of a system built for peacetime statistical reporting, and neither fully absorbs military deaths, occupied territory, or the composition of who is actually present. This site charts three Ukrainian series below and deliberately does not chart those two, because a line that says the wrong thing confidently is worse than no line.
The habit worth taking from this is not scepticism about Ukraine specifically. It is that wartime statistics from any country, including the ones you find sympathetic, are produced by institutions under stress and are frequently measuring a different thing from the thing named in the label. Where a number below is soft, it is said to be soft.
Why the state did not collapse
In February 2022 the expectation of rapid Ukrainian state collapse was close to universal, and it is worth being precise that this was not a Russian delusion that Western analysts saw through. Russian planning appears to have assumed it. So did most American and European intelligence assessments, which is why the initial Western offer to Ukraine's leadership was evacuation. Forecasters on all sides were wrong in the same direction at the same time, which is a useful warning about how much of that analysis was independent.
The interesting question is not why Ukrainians resisted, which every account can supply a motive for, but why the administration kept running - why pensions were paid, schools re-opened, trains ran on schedule, and the electricity grid was repaired faster than it was destroyed. The most persuasive structural answer is a reform that drew little attention outside Ukraine at the time. Between 2014 and 2020 the country amalgamated some eleven thousand local councils into roughly fifteen hundred larger communities and, critically, gave them a real share of national tax revenue and genuine control over how it was spent. The intent was ordinary public-administration efficiency. The effect, unplanned, was that when the centre was overwhelmed in 2022, the parts of the state that mattered day to day were not run from the centre and did not need instructions to keep operating.
Digital administration compounded this. Ukraine had built a national app, Diia, that put identity documents and a growing list of government services on a phone before the war; it turned out to work when the physical offices did not. Both of these are worth separating from morale, because morale is unfalsifiable and administrative structure is not. The general mechanism is transferable and slightly uncomfortable: a state's resilience under shock depends less on the quality of its central leadership than on how much of its routine function has already been pushed out to places the shock does not reach. That is not a claim that Ukraine's institutions are strong in general. On most peacetime measures they are not. It is a claim that decentralised administration failed gracefully, which is a different property from being good.
The economy that runs on other parliaments
Ukrainian output fell by roughly 29% in 2022. That is a larger single-year contraction than any developed economy recorded in the 2008 crisis or the pandemic, and it is the number the subsequent growth figures have to be read against. Output grew about 5.5% in 2023, 3.2% in 2024 and 1.8% in 2025 - real growth, and decelerating, from a floor far below where the economy started. A recovering percentage says nothing about the level it is recovering to.
The structural fact underneath is the one that gets least attention and constrains the most. Ukraine's finance ministry put external support required for 2026 at about $49bn, equal to roughly 43% of state budget spending, and when the budget was adopted something like $19bn of that had no committed source. The International Monetary Fund approved a new four-year facility of $8.1bn in February 2026, which is significant as a seal of approval and small relative to the requirement. Much of the rest comes through European Union facilities and a mechanism financed by the profits earned on immobilised Russian central bank assets - notably not by the assets themselves, a distinction covered on the dollar and sanctions pages and one that took European governments four years to arrive at.
State this without adjectives and it is simply a structural condition: a country whose budget is two-fifths externally financed has fiscal sovereignty that is a function of votes in foreign legislatures. It shapes what Ukraine can refuse in a negotiation, how much leverage donors have over domestic reform, and how quickly the whole arrangement could change. It is also the reason the corruption question below is not a moral sidebar but a live financial variable: the money is appropriated by parliaments that have to justify it to their own voters.
The military burden is the other half of the arithmetic. Defence spending moved from about 3% of GDP before 2022 to the mid-thirties, among the highest sustained military burdens recorded for any country in the modern era. The share is inflated from both directions - spending rose steeply while the economy it is measured against had shrunk - but no adjustment brings it near normal. A state spending a third of its output on defence and covering two-fifths of its budget from abroad is not running a wartime version of a normal economy. It is running a different kind of economy, and the transition back is a distinct problem from winning or ending the war.
Corruption, without either script
This is the subject on which the two information environments distort in opposite directions, which makes it the one most worth doing carefully. One account holds that Ukraine is irredeemably corrupt and that aid sent there is stolen. The other holds that raising corruption during a war is itself a hostile act, an argument that has been made in good faith by people worried about the aid and in bad faith by people who benefit from the silence. Both accounts fail against the same evidence, and the evidence is unusually checkable.
Take the measurable part first. On Transparency International's Corruption Perceptions Index, Ukraine scored 26 out of 100 in 2014, ranking 142nd of 175 countries. It scored 36 in 2025, ranking 104th of 182. That is the largest improvement of any current European Union candidate country over the period, and it is also a plateau: the score has moved between 35 and 36 since 2023 and gone nowhere. And 36 remains a poor score in absolute terms, well below any EU member state. All three of those statements are true simultaneously, and each of the two standard accounts is built by selecting one of them.
The institutional history is more informative than the score. After 2014 Ukraine built a set of anti-corruption bodies with genuine independence - the National Anti-Corruption Bureau (NABU), a specialised prosecutor's office (SAPO), and a High Anti-Corruption Court - partly out of domestic pressure and partly as a condition of Western money. In July 2025 parliament passed and the president signed a law subordinating NABU and SAPO to the prosecutor general, which would have ended their independence in substance. What happened next is the part worth remembering. Ukrainians held the largest street protests since the invasion began, in wartime, under martial law, during air raids. Within nine days the law was reversed and independence restored by a vote of 331 deputies.
Then the machinery that had just been saved was used. In November 2025 NABU and SAPO exposed a roughly $100m scheme centred on the state nuclear operator Energoatom, allegedly organised by a businessman who had been a close associate and former business partner of the president. Two serving ministers resigned within days. In February 2026 the former energy and justice minister was charged with money laundering and participation in organised crime, detained while attempting to leave the country, and refused release by the High Anti-Corruption Court; investigators said more than $7m had moved to foreign accounts naming his wife and children. Eight further suspects were charged and the investigation reached into fifteen jurisdictions.
Now read those two paragraphs against each other, because the sequence is the finding. The corruption was real, large, and located near the top. The attempt to switch off the bodies investigating it was also real, and came from the government itself. And the attempt failed, at a moment when the excuse for succeeding - wartime necessity - was as strong as it will ever be. What distinguishes a captured system from a contested one is not the absence of corruption, which no country achieves, but whether the machinery that exposes it survives being switched off. On that specific test Ukraine has a public, dated, documented result.
Two honest qualifications, in both directions. The result was contingent rather than structural: it depended on street mobilisation and on donor conditionality, and neither is permanent. Ukraine's anti-corruption officials have continued to describe pressure to slow investigations. Against that, the frequently-heard claim that Western military aid has been diverted at scale is a different claim from the domestic-procurement corruption that is documented. The audits conducted so far have not substantiated it. That is not the same as proof that it did not happen - the auditing of wartime aid is genuinely incomplete. Someone telling you confidently that the aid was stolen and someone telling you confidently that none of it was are both ahead of the evidence.
Governing without an election
Everything above describes a state that kept working. It does not describe who is authorised to run it, and that question has been open since 2024. Ukraine has held no national election since the invasion. Volodymyr Zelensky's five-year term expired on 20 May 2024, parliament's term has been extended in the same way, and martial law has been renewed in ninety-day increments continuously since February 2022. Any account of the country as a functioning system has to deal with that rather than around it.
The legal position is clearer than the political argument about it. Ukraine's constitution bars parliamentary elections while martial law is in force and extends parliament's mandate until a new assembly can be elected; the electoral code extends the bar to all elections; and a separate article provides that a president exercises the office until a successor is sworn in. On that reading nothing irregular has occurred: the constitution anticipated exactly this situation and specified what happens. Parliament restated the position formally, affirming that the president was elected in a legitimate election and that no vote can be held until martial law ends. The Russian government argues the opposite - that an expired term voids the authority to sign anything binding - which is a position with an obvious interest attached, since it would mean there is no counterparty capable of concluding a settlement.
The harder questions are practical rather than constitutional, and they are the reason the date keeps receding. Roughly a fifth of the country is occupied and its residents cannot vote. Around five and a half million citizens are outside the country and several million more are displaced inside it, all of whom would need a way to register and vote. Hundreds of thousands are serving in the armed forces. Campaigning requires public gatherings in a country under regular missile and drone attack. Ukrainian officials have said an election would need three to six months of preparation after martial law is lifted, and the electoral administration would have to be rebuilt for conditions no European democracy has run a vote in.
Martial law has also changed domestic politics in ways that are worth stating plainly rather than filing under wartime necessity. Several parties with Russian links were banned in 2022. National television news was consolidated into a single joint broadcast, which removed most editorial competition on the largest platform even as independent online outlets continued to operate. Opposition activity has been constrained by the same emergency framework that suspends the elections. Each measure has a security rationale that is not frivolous. Together they mean that the normal mechanisms for testing a government - a scheduled vote, a competitive media market, an opposition with a route to power - have been substantially unavailable for four years, in a country whose anti-corruption institutions the previous section described as saved by street pressure rather than by constitutional insulation.
What polling exists suggests the eventual contest will be real rather than a formality. Valerii Zaluzhnyi was commander-in-chief until his dismissal in February 2024, and has been ambassador to the United Kingdom since. He records the highest trust of any public figure in Ukraine, around 70% in Rating Group surveys. Hypothetical ballots put Zelensky first in a crowded field at roughly a quarter of the vote, with Zaluzhnyi second, and give Zaluzhnyi the advantage in a two-round runoff. These numbers deserve the usual wartime caveats - a population under attack, an incomplete electorate, and no campaign to test them against - but they do indicate a public that distinguishes between supporting the war effort and endorsing the incumbent.
For the paths below, this matters in a specific way. Any negotiated settlement would have to be accepted by a Ukrainian government, and the sharpest domestic argument about any settlement is likely to happen after martial law ends rather than before. A country that has not voted in four or more years, with a returning diaspora, an army being demobilised and a set of painful compromises to ratify, is entering the hardest phase of its politics rather than leaving it. The administrative resilience described earlier is the thing most likely to be tested by that, and it is a different test from the one it has passed so far.
Who is left, and who comes back
About 5.6 million Ukrainians were still living abroad as refugees at the start of 2026, and roughly 1.85 million had returned since the initial exodus. Surveys by the Centre for Economic Strategy, a Kyiv think tank that has tracked the refugee population in repeated waves, find around 43% intending to return and about 36% saying they will not or probably will not. The composition matters more than the totals: adult women and children make up the large majority of those abroad, and return rates are lowest precisely among families whose home regions are occupied or still being fought over.
The uncomfortable finding from every previous large displacement is that stated intention to return decays with time, and decays fastest once children are enrolled in schools in the host country and a parent has a job there. Four years is already long by that standard. This is not a prediction that few will come back; it is a warning that the intention figures are the most optimistic version of the answer, and that the answer is being decided now by ordinary domestic decisions rather than later by policy.
Births tell the same story from the other end. Ukraine's fertility rate fell from about 1.15 births per woman in 2021 to below one in 2022 and has stayed there. Read that figure carefully, though, because it is calculated over the women actually present in government-controlled territory, and young women are heavily over-represented among those who left. It is measuring something different from a peacetime fertility rate. What it does establish is that the demographic hole is not only a matter of who emigrated; the cohort that would normally have been born during these years largely was not.
This is why demography, not territory, is the variable that determines what Ukraine is on the other side of the war. A country can rebuild a bridge. It cannot rebuild a missing cohort, and the reconstruction figures below assume a population to rebuild for.
What rebuilding would cost
The most careful estimate is produced jointly by the World Bank, the United Nations, the European Commission and the Ukrainian government. That matters: a joint assessment by parties with different incentives is harder to slant than any of them alone. Its February 2026 edition put direct physical damage at about $195bn as of the end of 2025. Broader economic losses came to about $667bn, and recovery and reconstruction over the following decade to roughly $588bn.
That last figure is close to three times Ukraine's annual economic output. No country has ever financed reconstruction on that ratio from its own resources, and the arithmetic is not close enough for growth to solve it. The largest single components are transport, at more than $96bn, and the power and heating systems, at about $91bn - the latter reflecting a sustained campaign against the grid, and also the fact that it has been repeatedly repaired, which is itself part of the cost.
Which makes reconstruction a political question rather than an engineering one, and a question with an unresolved answer: who pays, on what conditions, and with what control over the spending. The frozen-assets episode covered on the dollar page is the visible edge of it. The conditionality attached to reconstruction money is the same lever that produced Ukraine's anti-corruption institutions, and the same lever that makes their independence contingent on external interest that may not last. Every serious version of the reconstruction question runs back through the two sections above.
How Ukraine's strengths and weaknesses compare
The same exercise this site applies to Russia, applied here, and with the same rule: strengths and weaknesses are worth looking at separately rather than averaged into a verdict.
The paths from here
These are futures for the country rather than outcomes of the fighting, which is a deliberate distinction: several of them are compatible with more than one way the war ends.
The armed, externally financed state
Fighting subsides into a hardened line without a settlement. Ukraine remains heavily militarised, keeps defence spending far above normal indefinitely, and continues to cover a large share of its budget from abroad. Reconstruction happens in the safer west and centre while the east stays a frontier zone.
Will it happen? This is the base case on current trajectory, and the most important thing about it is that it is not stable in the way "frozen" suggests. It requires donors to keep appropriating money annually for a war that is no longer producing news, which is the hardest kind of spending for any legislature to sustain.
A settlement with credible guarantees
A negotiated end that includes security arrangements Ukraine and its backers regard as enforceable. Reconstruction money arrives at scale because the risk premium falls. Refugee return accelerates because the safety question is answered.
Will it happen? The two blocking questions have been stable for years - which territory, and what guarantee - and 2026 supplied three rounds of talks and two short truces without moving either. The guarantee problem is the harder one, because any guarantee strong enough to reassure Kyiv is close to the arrangement Moscow says it went to war to prevent.
Financing failure before military failure
The binding constraint turns out to be foreign appropriations rather than the front line. A change of government among major donors, or simple fatigue, opens a gap that cannot be closed domestically. Austerity, arrears, and a weakened negotiating position follow without any battlefield collapse.
Will it happen? This is the risk that receives least attention and the one Ukrainian officials privately rank highest. A $19bn hole in an adopted budget is not a hypothetical. The mechanism does not require anyone to decide to abandon Ukraine; it only requires several parliaments to be slow at the same time.
Accession as the organising project
European Union membership becomes the frame that disciplines domestic reform the way it did for Poland and the Baltics, with the conditionality doing the work that domestic politics alone would not. Institutions harden because the reward for hardening them is concrete.
Will it happen? Partly, and slower than either side of the debate suggests. The precedent is genuine and the domestic consensus is real. The obstacles are on the European side: Ukraine would be a large, poor, agricultural member, and existing members have not yet had the honest argument about what that costs them. Enlargement has never been blocked by a candidate's enthusiasm.
Reconstruction captured
The money arrives and a meaningful share of it is captured, either by domestic networks or through procurement structures that favour donor-country firms. Physical rebuilding happens; the institutional gains of the previous decade are eroded by the sheer volume of contracts.
Will it happen? Partially, almost certainly - this happens in every large reconstruction, including ones run entirely by wealthy democracies. The open question is degree, and the July 2025 episode is the reason to think the answer is contested rather than settled in advance. Watch whether the anti-corruption bodies retain independence once the conditionality that created them relaxes.
Rebuilt for people who do not return
Physical reconstruction proceeds while the demographic assumption underneath it fails. Housing, schools and hospitals are restored in regions whose populations have permanently relocated, in Ukraine or abroad. The country ends up smaller and older than any planning document assumes.
Will it happen? Some version of it is close to unavoidable given the fertility and displacement figures; the question is scale. It is the least discussed of these paths because it is nobody's argument - it does not support any position anyone is currently taking - which is usually a sign that a scenario is being underweighted rather than that it is unlikely.
Where serious analysts disagree
Whether the anti-corruption institutions are durable. One reading treats the July 2025 reversal as proof that Ukrainian civil society is now a genuine constraint on the executive. The other treats it as proof of the opposite - that the institutions needed rescuing at all, and were rescued by external pressure that will fade. Both readings fit the same events, and the disagreement is really about whether donor conditionality builds institutions or only rents them. It is testable, and the test arrives when the conditionality relaxes.
Whether refugee return is mainly about safety or mainly about time. If safety is the dominant variable, a settlement produces substantial return. If elapsed time and integration abroad dominate, a settlement arrives too late for a large share of those who left. The evidence from earlier displacements favours the second, and Ukrainian and European policy is largely built on the first.
Whether the war economy is a distortion or a foundation. One camp sees a defence-industrial sector built at emergency speed as an asset that can anchor an export industry and a NATO-adjacent role. Another sees a third of output in a sector with no civilian market and an economy that will face a painful conversion. Israel and South Korea are cited as evidence for the first; most post-war economies are evidence for the second.
What the population actually is. Noted above, and worth restating as a live disagreement rather than a technicality, because a difference between 29 and 35 million changes per-capita output, pension arithmetic, mobilisation capacity, and reconstruction planning simultaneously. Analysts who quote these figures without a range are usually not aware of how wide it is.
What this means for reading everything else
This page avoids the fighting for two reasons. It is the part everyone already covers, and it is the part where the reader can check least. The material above is checkable instead. Budget documents are published. Index scores are published. Court proceedings happen in public. And assessments produced jointly by parties with different interests are harder to bend than any single account.
That is the general point worth carrying away. When a subject is saturated with claims you cannot verify, the productive move is usually not to work harder at adjudicating them but to change which questions you are asking - toward the ones where evidence exists and where the answer constrains the others. How a state is administered, what it lives on, and who is left in it will shape Ukraine's next twenty years more than any particular line on a map, and unlike the map, they can be looked up.
The mechanics behind this
The forces at work on this page are explained from first principles elsewhere on the site. Decentralised administration failing gracefully is an institutional-design property; the conditionality that built Ukraine's anti-corruption bodies is an incentive structure; and the reason both propaganda accounts are internally consistent while contradicting each other is a matter of how motivated reasoning assembles evidence.