Fair deal: do plea agreements really work in top corruption cases

Debates still rage around the plea agreement in Serhii Pashynskyi’s case, which in turn raise a wave of new questions about this type of verdict. Let’s try to sort through some of them.

23 September 2026, the High Anti-Corruption Court (HACC) approved a plea agreement with former Acting Head of the Presidential Administration Serhii Pashynskyi in the case of the appropriation of arrested “Kurchenko’s petroleum products.” From the verdict it follows that Pashynskyi fully admitted guilt in abuse of official position and participation in a criminal organization and agreed to provide incriminating testimony. And the total amount he must pay or transfer for the benefit of the state and the Defense Forces is about 1,1 billion UAH.

Public reaction to the first report about this agreement was almost unanimous: “he bought his way out.” And this is not the first such reaction to plea-based verdicts.

In recent years, plea-based verdicts in top corruption cases have become more frequent. According to HACC data, overall from 2019 until the end of the first half of 2026 the court issued 201 verdicts based on plea agreements. For the first four years these were isolated cases — from one to ten verdicts per year. But then a sharp rise began: 27 verdicts in 2023, 38 in 2024 and as many as 72 in 2025.

The most recent jump in numbers does not surprise us, since it coincided with the first full year of the law that optimized concluding agreements in corruption and corruption-related cases. And first-half data for 2026 (35 verdicts) indicate that the plea agreement institution has reached a relatively stable level. But does it really work?

How plea agreements are concluded in Ukraine

We’ll remind you up front that a plea agreement is a criminal-law compromise. The state agrees to a milder punishment, and the defendant in return admits guilt and takes on several obligations: to expose accomplices, to compensate damages, and to accept an agreed punishment. In other words, an agreement is about convicting a person and recognizing them as guilty.

The content of the agreement is set by the parties — the prosecutor and the defendant together with defense counsel. The court approves the agreement — in top corruption cases that is the HACC — and does so by issuing a conviction. Therefore a person who entered into an agreement is convicted, not “released from responsibility.”

Upon receiving the indictment together with the agreement, the court checks whether the agreement’s terms comply with the law and the public interest and whether they violate the rights of other persons. The court cannot rewrite the agreement or reallocate payments provided by it — it can only approve the document or refuse to approve it, explaining why it conflicts with the law.

Refusals to approve agreements at the pre-trial stage are not common; as of mid-September HACC reports only four cases where agreements were not approved and the case was sent for pre-trial investigation. For example, in the “Kurchenko petroleum products” case the court previously refused to approve an agreement with Serhii Pashynskyi, finding that it provided punishments not permitted by law, did not ensure full compensation of damages, and placed performance of the agreement on an indeterminate circle of third parties.

From the start there were warnings about agreements in corruption cases. The main one is the risk that an agreement could become a way to “make a deal” with those organizers of schemes who are most responsible for the crimes. Therefore the law limited the circle of persons and grounds for agreements. The new law on agreements, adopted in October 2024, expanded the grounds: now an agreement can be concluded if the defendant exposes another person for committing any corruption or corruption-related crime, and also on condition of compensation for damages and the defendant’s exposure of other suspects.

At the same time, the law substantially increased the incentive for defendants. It introduced article 69–2 of the Criminal Code of Ukraine (CCU), which allows imposing a sentence below the statutory minimum where other persons are exposed. If the court imposes such a sentence, the person cannot then be released from serving it on probation.

The upper limit of imprisonment for which release on probation is possible under an agreement was increased from 5 to 8 years. It also became possible to impose an additional fine even when the article’s sanction does not provide for it. The procedure changed as well. Agreements are approved by higher-level prosecutors depending on the severity of the crime. Repeatedly approaching the court with an agreement with the same person is allowed only once and only after the reasons for an earlier refusal have been eliminated. This is exactly what we saw in the case of the repeat plea agreement with Serhii Pashynskyi.

But are these updates enough for genuinely high-quality agreements and how have they worked so far?

On the one hand, for the first time there have been verdicts with punishment below the statutory minimum under article 69–2 of the CCU. In at least three such cases the convicted persons still serve real, though shorter, prison terms. These are former Supreme Court head Vsevolod Knyaziev, former Prosecutor of the Prosecutor General’s Office Volodymyr Dergunov, and former Ministry of Defense official Oleksii Stasiuk. On the other hand, the additional fine under part 3 of article 53 of the CCU is applied, albeit not always, as is special confiscation.

The cooperation/exposure mechanism is also producing results. According to the Specialized Anti-Corruption Prosecutor’s Office (SAPO), on the basis of testimony from convicted persons who agreed to cooperate, 15 people have already obtained verdicts, and in most proceedings where such testimony has been or will be provided, trials are ongoing.

Regarding special confiscation, the law still did not directly establish rules for its mandatory application when concluding agreements, although we recommended such a change. In our view this is a gap that needs closing: if there are grounds for special confiscation, the court should approve the agreement only if it provides for it. That way the state would certainly obtain property acquired as a result of the crime.

This position is supported and supplemented by authors of the judicial practice analysis for 2024–2025 years, prepared by the Center for Political and Legal Reforms together with the National Agency for Prevention of Corruption (NAPC). They specifically note that courts often lifted arrests on land plots, vehicles and funds precisely because the agreement did not include a confiscation requirement. In other words, property that the state had already blocked was returned to convicts because of the particular wording in the agreement.

At the same time, a telling example is the court’s refusal to approve an agreement with former MP Oleksandr Kachnyi in the illegal enrichment case. The agreement provided for special confiscation of an apartment, a house and land that were formally owned by the husband of a deceased nominal owner. Although the man had acted under a power of attorney when buying the apartment, inherited the property for free, and in court directly confirmed that he was only the nominal owner, the court refused to approve. The reason was the allegedly unproven awareness by the third person of the illegal origin of the funds. In our view, those circumstances did not justify such doubts.

Main problems with current agreements

Despite the many benefits of agreements, society does not always view them favorably, because the state does not always explain such verdicts well. Here are, in our view, the main reasons why that happens.

The first reason — sealed verdicts.

The texts of some plea-based verdicts, including those in Pashynskyi’s case, former head of the Western Appellate Commercial Court Borys Plotnitskyi and other defendants, are fully sealed. Because of this, from the court’s and SAPO’s public communications it is impossible to learn what exactly the state obtained in exchange for the mitigation of punishment. And since these are often well-known people, public outrage becomes an expected reaction to such secrecy.

One of the earliest and still most high-profile sealed verdicts was the agreement with former Ecology Minister Mykola Zlochevskyi, who tried to close the criminal case against him by offering a “superbribe” of 5 million USD to the heads of NABU and SAPO. Under the plea agreement concluded in August 2023, HACC fined Zlochevskyi 68 thousand UAH. In addition, under the agreement he donated 660 million UAH to the Armed Forces. And that was it. Because the verdict was completely classified, the public never learned whether the “public interest” rules were followed and whether the state truly benefited from those arrangements.

That agreement in fact became one of the main reasons to talk about updating the legislation on such verdicts. Since then, we at TI Ukraine have consistently emphasized the need to publish the maximum permissible amount of information about an agreement so that Ukrainians do not have to raise such questions.

Full sealing often fails even at its direct function — hiding cooperation with investigators. After all, if by law a convicted person is obliged to testify against accomplices, those accomplices will still learn about the agreement through the mere fact of the separation of proceedings.

We’ll recall that HACC already knows how to redact specific parts of verdicts without sealing the entire text. For example, in the case of the head of the Agency for Local Roads of Poltava region the court left public the justification of the public interest and other parts of the verdict. And that same verdict in the Pashynskyi agreement was initially sealed, which we immediately flagged as a rather bad precedent. HACC quickly responded to our criticism and published a partially redacted verdict from which the details of the agreement became clear.

The second reason — lack of clear rules for concluding agreements.

In response to our request SAPO confirmed that there is no separate internal methodology for assessing the possibility of concluding an agreement. At the same time, the Office’s development strategy for 2026–2028 foresees drafting unified approaches — prosecutors plan to involve judges, lawyers and civil society organizations in that development.

HACC also told us that it had not systematized practices on agreements. Meanwhile a discussion about finding a balance between the efficiency of criminal justice, inevitability of accountability, fairness of punishment and public trust in the anti-corruption system is already underway. And the preconditions for developing clear policies on concluding agreements are in place.

Pashynskyi’s agreement well illustrates this process. In SAPO’s statement they listed the amounts of fines and compensations, but there were no explanations why these conditions benefit the state. The necessary explanation appeared only after the verdict was published. Thus we saw that the prosecutor, within the agreement, obtained testimony regarding other participants, and overall the number of defendants in the case decreased, which gives a chance to consider the cases against the remaining defendants before the statute of limitations expires.

However that information appeared almost a week after the news about the agreement, by which time society had already been discussing and outraged over the “billion for freedom” for several days.

The fact that compensation to the budget (600 million UAH) is less than the established damage (817 million UAH) also turned out to be not accidental. Pashynskyi’s case concerns a criminal organization, not a single defendant, and the damage was caused by the actions of seven people, not one. Therefore the parties agreed on a partial release of Pashynskyi from civil liability, taking into account the obligations of other convicts under agreements (111 million UAH in compensation and 40 million UAH to the Defense Forces) and his own 500 million UAH for army needs. Overall, compensation to the budget from all convicts in this case does not yet reach the established damage. Whether future verdicts regarding the remaining defendants will cover it is a question for SAPO.

The third reason — public expectations. And the public wants punishments that feel just.

Last year a TI Ukraine sociological survey showed that 60% of Ukrainians consider the most demonstrative criterion of effective anti-corruption efforts to be the number of actually punished corrupt officials, and such punishment should involve imprisonment or confiscation.

Agreements imposing fines and donations to charity do not meet that expectation. Therefore SAPO must explain them especially thoroughly, and the rules by which prosecutors agree to make deals should be clear before the next high-profile case appears.

Are the funds honest?

In almost half of the verdicts we analyzed that were concluded between 1 July 2025 and 30 June 2026, defendants committed to make charitable contributions for the needs of the Defense Forces. This practice appeared after Russia’s full-scale invasion of Ukraine, that is, before the updated law on agreements in top corruption cases was adopted, but today it can be considered almost mandatory for this type of verdict.

Amounts of charitable contributions range from 10 thousand to 20 million UAH. According to HACC, in 2025 alone over 321 million UAH were directed to the Armed Forces. And according to SAPO’s latest data, thanks to agreements in top corruption cases more than 1,6 billion UAH have already been directed to the Defense Forces. Of these, over 1,1 billion UAH came through the state fundraising platform United24, more than 272 million through the Return Alive foundation and over 50 million UAH through the “Sternenko Community.”

Pashynskyi’s agreement became one of the record sums mentioned in it: 600 million UAH — to compensate damages to the state budget in stages over 2,5 years, 400 million UAH or weapons worth that sum — to the Return Alive foundation within six months after the verdict, 100 million UAH — to the “Sternenko Community” for weapons within 30 days, as well as 850 thousand UAH fine as the main punishment and 17 thousand UAH fine and a one-year ban on holding public office as an additional penalty.

Three questions arise around these funds, and in fact they concern all the amounts mentioned in other plea-based verdicts.

1. Has inflation “eaten” the compensation?

One of the most common arguments from critics of Pashynskyi’s agreement is that damages caused in 2014 are being compensated today by a nominal sum that has significantly devalued over twelve years. In other words, 600 million UAH in 2026 is not the same money that Ukraine lost in 2014 due to the schemes of the former Acting Head of the Presidential Administration.

The argument has merit, but it is directed not at plea agreements as an institution, but at how the state generally protects its property interests in criminal proceedings.

Legislation does not generally leave the state defenseless against inflation. Under part 2 of article 625 of the Civil Code of Ukraine (CCU), a debtor who delayed a monetary obligation, at the creditor’s request, pays the debt taking into account the inflation index and 3% per annum. The Grand Chamber of the Supreme Court as early as 2018 stated that this rule applies to all kinds of monetary obligations regardless of the grounds of their origin, i.e., it should apply to damages awarded in criminal cases. In addition, article 1192 of the CCU provides that the size of damages for loss of property is calculated at its actual value at the time of the court hearing.

For a case where the subject is appropriated petroleum products, as in Pashynskyi’s case, this can have practical significance, since it concerns property whose price changed with the market and not a fixed monetary sum. However, to recover that difference it would most likely have been necessary to carry out another expert examination, which would also take time and might not lead to any result.

Claims to recover inflationary losses can be made in criminal proceedings by filing a civil claim. Case law knows such examples. In a case of embezzlement (part 4 of art. 191 CCU) Kamianets-Podilskyi City District Court recovered inflationary losses and 3% per annum from the defendant. However, because of the statute of limitations defense, the claim concerned only interest for the last three years, so instead of the claimed more than 419 thousand UAH in inflationary losses the defendant had to pay the state about 30 thousand.

There is such a case in HACC as well. In the case of the head of a Defense Forces department Oleh Halushka, the Ministry of Defense claimed almost 8 million UAH in inflationary losses from more than 21,5 million UAH of primary damage. However, the court issued an acquittal and left the claim without consideration. The review of that verdict is still ongoing — in June 2026 the Supreme Court reversed the APU HACC decision confirming the legality of the acquittal, so the issue of compensation including inflationary losses may be revisited.

To recover inflationary losses the obligation must be monetary, i.e., clearly defined in a specific sum. In criminal cases this is usually possible only when it concerns embezzlement or appropriation of funds, not other types of damage. In Pashynskyi’s case it was not about seizing a certain sum of money — the subject was petroleum products.

So the problem is not that the law prohibits accounting for inflation, but that no mechanism makes it automatic. A court cannot independently index the size of damage and go beyond the claims submitted. And if the injured state body or the prosecutor in the state’s interest did not claim inflationary losses, they will not be recovered either in a conviction in ordinary proceedings or in proceedings based on an agreement.

In Pashynskyi’s case no civil claim was filed at all. Therefore a claim for inflationary losses could not have been granted by the verdict nor reflected in the agreement, and the size of the damage was fixed in nominal terms as of 2014 — 600 million UAH. Moreover, even a filed claim for such compensation faces the risk of the statute of limitations, which in cases involving events a decade old can reduce it to symbolic sums.

Hence the recommendation. The issue of inflationary losses should be addressed primarily to the prosecution (in our case — SAPO) and to the bodies representing the injured state. A claim for their recovery should become a standard part of civil claims in cases of theft of state property, and its inclusion should be one of the criteria by which a prosecutor assesses the amount of compensation in an agreement.

Separately one should discuss whether the requirement of “full compensation of damages” as a precondition for an agreement covers only the nominal sum of damage or also its real value at the time of concluding the agreement. Today the law does not directly answer that question.

Additionally, part of the public’s complaints about the “leniency” of punishment concerns issues that an agreement does not resolve. Confiscation of property in Pashynskyi’s case was not imposed: SAPO explained that the relevant additional punishment under article 255 CCU was introduced in June 2020, i.e., after the events of the case. So here the non-retroactivity principle applies — a law should not worsen a person’s situation retroactively.

2. Why do funds go to foundations and not to the budget?

Former HACC head Vira Mykhailenko once explained that the prosecution and defense agree where exactly to direct charitable payments and in what amount, and the court only checks whether the agreement complies with the law. So the question “Why these particular funds?” should be addressed to the parties to the agreement, not to the court.

However, the question should be asked honestly. Are there justified concerns about the transparency of the operations of the funds mentioned in the agreement? And if there are no such concerns, what is the difference for the state between funds going to the budget and funds going to a foundation that professionally and accountably procures weapons for the army?

In wartime a large portion of the budget is already directed to defense. A foundation — Return Alive, the Sternenko Community or any other — can be a faster and more targeted channel in that chain. Moreover, Return Alive expressly states that it does not take a percentage of donations for administrative costs.

The real risks are different. The first is that a fund is not connected to the defendant or their circle. The second is that the fulfillment of the condition can be verified.

Non-performance of the agreement carries separate criminal liability, and SAPO may ask HACC to cancel the verdict and consider the case in the ordinary manner. Notably, during the court’s operation prosecutors have never filed such a motion.

According to SAPO, there have been no detected cases of deliberate non-performance of agreements so far, and no convicted person has been held liable for breaching agreement conditions. Performance is confirmed by payment documents that are provided to the court and prosecutor and attached to the case file. So verification of fund transfers exists, but it is done solely by SAPO. Beyond that one can rely on the public reporting of the relevant charitable foundations, which should be open to the general public. If there is a demand for reporting specifically from HACC plea agreements, it would make sense for those foundations to publish more specific information about what exactly the funds received under such agreements were spent on. That would help to understand what concrete needs of the army are covered through such donations.

At the same time, the public reaction to Pashynskyi’s obligation under the agreement to donate to the Return Alive and Sternenko Community funds also produced unexpected results. 28 September HACC did not approve the plea agreement of MP Oleksandr Kachnyi in the illegal enrichment case for over 12,7 million UAH. Kachnyi admitted guilt and agreed to transfer 3 million UAH to the Return Alive project. The judge in his refusal stated that the use of those funds would be beyond the control of the parties and the court, and therefore the court could not be convinced that those provisions correspond to the public interest.

3. Whose money is it?

This is the third and perhaps the most important risk. A formally positive outcome — paid compensation or a donation to the army — can be ensured by opaque money. The origin of funds used to fulfill an agreement’s conditions should be checked no less carefully than, for example, the origin of bail funds.

SAPO confirmed that in practice both convicted persons and third parties reimburse damages and make charitable contributions — the law does not prohibit this. At the same time, there is no separate procedure by which a prosecutor would verify the origin of such funds in the Criminal Procedure Code. The Office relies on banks, which as subjects of primary financial monitoring must check transactions in accordance with anti-money-laundering legislation.

However, one should be consistent and acknowledge an uncomfortable point for critics — namely, to recall the special confiscation mechanism. After all, special confiscation is about seizing precisely those funds and assets whose criminal origin is directly established by a court verdict. When it is applied, the state deliberately fills the budget with money obtained by criminal means.

So purely pragmatically it is difficult to explain why the same logic becomes unacceptable when funds come not through special confiscation but as compensation for damages or as charitable terms of an agreement. In the end the money ends up in the budget or the army and no longer remains with the convict. Thus the outrage that “the money didn’t go to the budget” concerns form, not substance.

Why agreements are still good, and why they should be improved

The main argument in favor of Pashynskyi’s agreement is not its amount but the alternative. The events of the case took place in spring 2014, while suspicions were only notified in February 2024. The statute of limitations for abuse of office in this case expires already in 2028, so the parties directly linked the agreement to making it possible to try the case against other defendants within that statute.

A full trial with multiple defendants, a large body of evidence and defense procedural tactics could have lasted years. If the statutes of limitations had expired in the meantime, the person would have been released from liability entirely, and the state would have obtained nothing: no verdict, no compensation, not even partial. Despite European Commission demands, the rules on limitation periods in Ukraine have still not been reformed. So in cases investigated decades after the events, a plea agreement is often not an alternative to a harsh sentence but an alternative to impunity.

Criminal proceedings always have a probabilistic result. The prosecution believes guilt is proven, the defense that it is not, and the court decides. An agreement turns probability into a guarantee: there is a conviction, the status of a convicted person, real money and testimony against other participants. A fine is a form of primary punishment under the Criminal Code. So to talk about “buying off” a punishment that is itself a punishment is incorrect.

For comparison, another defendant in Pashynskyi’s case, Serhii Tyshchenko, under an agreement received eight years’ imprisonment with release on probation, transferred 40 million UAH to the Defense Forces and must pay an additional 110 million UAH to the budget over three years. The “Kurchenko petroleum products” case is not finished: there were other suspects, and not all of them took agreements.

At the same time, precisely because agreements have become the main way to conclude top corruption cases, their quality must continue to be improved. HACC should systematize approaches to assessing the public interest and make partial redaction of verdicts a standard practice instead of full sealing where full secrecy is not required for protection of the person or other proceedings. SAPO should implement measures in its Strategy on unified approaches to agreements and publicly explain in each high-profile case how the state benefits.

Parliament needs to make special confiscation a mandatory condition for approving an agreement where there are legal grounds. Broader use of real, though shorter, prison terms and additional fines would also improve public perception of such verdicts.

And finally, two problems that most fuel outrage about agreements lie entirely outside them. These are unreformed statutes of limitations, which force prosecutors to negotiate, and the fact that the state does not always use existing mechanisms to recover inflationary losses, which makes nominal compensation devalued — whether under an agreement or a regular verdict. After all, hundreds of millions of hryvnias of damage caused in 2014 are not the same as the same hundreds of millions in 2026.

These problems must be solved by both legislators and practitioners. And until they are addressed, an honest plea agreement remains a better result than a case that ends with nothing.