GAB is pleased to reprint the article below from Vox Ukraine Idea, an independent analytical platform dedicated to helping Ukraine move into the future. Authors Professor Yuriy Gorodnichenko of the University of California, Berkeley, and Ilona Sologoub, Vox Ukraine’s Scientific Editor, have taken a major step forward in explaining how policymakers can manage the vexed and misunderstood issue of corruption perception measures.

In March 2026, Ukrainians reported that corruption was the second most important problem after the war. 12% of people even put it in the first place. Between  70% and 90% of Ukrainians believe that corruption is a serious problem. At the same time, the incidence of corruption was much lower: in 2025, between 5% (in administrative services) and 32% (in the construction sector) of people found themselves in situations where a bribe was necessary to address their issues. This divergence between perceived and experienced corruption has been persistent: the gap has reached 60-70 percentage points at least since the early 2000s, when the data became available. 

This situation is not unique to Ukraine. In many countries, perceived corruption differs from experienced corruption. Furthermore, when objective measures of corruption are available (Sarullo et al. 2026), they are only weakly correlated with perceptions. One explanation is that surveys measure only petty corruption, whereas perceptions of grand corruption are shaped by the media. Consistent with this explanation, freedom of speech is related to perceived corruption (Gutmann et al. 2020, Costa 2013). This can lead to the integrity paradox: more information about officials prosecuted for corruption can increase popular beliefs about the extent of corruption in a country. 

How can one then defeat the corruption narrative?

A natural first step is to communicate anticorruption progress and therefore rehabilitate Ukraine’s reputation. But given the integrity paradox, the net effect of communication can be attenuated or even turn negative. So what is the net effect? To address this key question, our recent work  (Gorodnichenko et al. 2026) tests directly how information about the incidence of corruption and about prosecution of corrupt officials affects people’s corruption perceptions.

Specifically, we implement a randomised controlled trial for a sample of over 7000 Ukrainians. The first wave of the survey experiment was fielded in February 2024. A follow-up wave was fielded in April 2024. The sample structure corresponds to that of the Ukrainian population as of January 2022. We split the sample into one control and four treatment groups. The control group was provided with neutral information, while treatment groups were informed about various aspects of corruption and anti-corruption efforts (Table 1).  

Table 1. Information interventions

Group	Provided information
Control group	Before you proceed, we would like you to know that 75% of Ukrainians are preparing for blackouts in the winter, and 11% plan to move to another house within their settlement.
Treatment T1	Before you proceed, we would like you to know that since July 2021 the high anticorruption court convicted 75 high-level officials.
Treatment T2	Before you proceed, we would like you to know that in 2022 35% of firms paid a bribe at the customs office and 19% of firms that interacted with the judiciary system were involved in corruption.
Treatment T3	Before you proceed, we would like you to know that since July 2021 the high anticorruption court convicted 75 high-level officials. We would also like you to know that in 2022 35% of firms paid a bribe at the customs office and 19% of firms that interacted with the judiciary system were involved in corruption.
Treatment T4	Before you proceed, we would like you to know that in February 2023 the Supreme Court confirmed the conviction of Hennadii Diachenko and Ivan Radyk who diverted UAH 116 million intended to build a rail link between Boryspil airport and Kyiv. They will serve 11 and 10.5 years in jail.

 Before and after the treatment we ask respondents about their perceptions of corruption in general and by agency (e.g., President, Parliament, local government, healthcare, etc.). We also ask respondents about their perceptions of the government’s willingness to fight corruption and effectiveness in doing so. A number of supplementary questions elicit media consumption, personal corruption experience, tolerance to corruption, and knowledge of specialized anti-corruption institutions established in Ukraine since 2014. To see how corruption perceptions impact actual behaviour, we asked whether people were willing to donate or volunteer in response to requests from different actors (government, army, volunteer, friends, etc.), and about their emigration intentions. In the second wave of the survey we asked about corruption perceptions and whether respondents donated or volunteered during the last month.

Basic facts

Consistent with other survey evidence, Figure 1 shows that corruption is broadly seen as a pervasive, major issue in Ukraine. Almost 50% of respondents assign the highest score (10) to the prevalence of corruption in the country. Few respondents assign a score of less than 5. In addition, more than 80% of respondents believe that corruption is the most important problem (score of 10) for the country. 

These poor scores correlate with actual behavior. For example, people with higher perceived corruption are less likely to donate or volunteer in response to requests by central or local governments. However, willingness to donate or volunteer when nudged by friends, volunteers or the army is not affected. Perceived corruption also reduces willingness to stay in Ukraine.

More basic facts can be found in this article

Figure 1. Responses to the questions on corruption perceptions in Ukraine

We also find that media consumption affects perceptions of corruption: people for whom TV is the primary news source have lower corruption perceptions than those who rely on online media or social media (note: since 2019 social media overtook TV as the major news source; since 2022 TV is largely controlled by the government with a United Telethon). Interestingly, knowledge about anti-corruption institutions does not correlate with perceived government willingness to fight corruption. It does, however, affect perceptions of corruption in certain government agencies. Personal experience of corruption in a specific area (e.g., paying a bribe at a hospital) is associated with higher perception of corruption in that area (in this example, healthcare).

Causal results

While correlations are informative, they do not entail causal relationships. To address this concern, we now utilize our randomized controlled trial (RCT) to establish causal effects of news about law enforcement on perceptions of corruption.  

The RCT results (Table 2) show that our treatments do not affect people’s corruption perceptions. This can be interpreted as either the support for the “integrity paradox” among Ukrainians or as the fact that our information treatments cannot change the extreme beliefs of the majority of respondents (Figure 1). However, treatments (specifically, providing information on convictions) do affect Ukrainians’ beliefs about the government’s willingness to fight corruption. Provision of the general statistics has a stronger effect than information on a single high-level case. However, in the follow-up wave the effect of a single case becomes more significant, thus suggesting that this specific information is more easily remembered.

We also find that information treatments do not affect either people’s willingness to donate or volunteer, nor their actual behaviour. Perhaps people believe that they should support the army regardless of the government’s quality.

Table 2. Treatment effects on measures of overall corruption and will to fight corruption. 

	Immediately after the treatment			Two months after the treatment
	Importance of corruption	Will to fight corruption		Importance of corruption	Will to fight corruption
	(1)	(2)		(3)	(4)
Treatment
T1 (convictions)	0.051	0.323***		-0.001	0.231
	(0.061)	(0.106)		(0.086)	(0.159)
T2 (incidence)	-0.019	-0.078		0.030	0.171
	(0.067)	(0.105)		(0.086)	(0.156)
T3 (convictions & incidence)	0.013	0.172*		0.013	0.440**
	(0.065)	(0.103)		(0.081)	(0.171)
T4 (high-level case – conviction)	0.010	0.189*		-0.130	0.396***
	(0.073)	(0.104)		(0.109)	(0.152)
Pre-treatment beliefs
Effectiveness of anticorruption agencies(1=min, 10=max)	-0.002	0.508***		0.000	0.352***
	(0.012)	(0.020)		(0.015)	(0.030)
Prevalence of corruption(1=min, 10=max)	0.304***	-0.264***		0.283***	-0.303***
	(0.020)	(0.021)		(0.026)	(0.031)
Observations	7,324	7,324		4,888	4,888
R-squared	0.158	0.302		0.138	0.179

Conclusion

Simply informing citizens about the number of prosecuted officials is unlikely to reduce their corruption perceptions. If anything, this can strengthen their view that “there are many corrupt officials out there”. At the same time, informing about the enforcement does strengthen people’s beliefs about the government’s will to fight corruption. This calls for persistent anticorruption efforts to eventually overcome the entrenched perceptions. This also underscores the importance of putting the success of anticorruption efforts in a broader context so that the public can appreciate the progress.

We also suggest that Ukraine’s international partners refrain from general statements about a “very corrupt Ukraine” but instead focus on specific matters such as support for anti-corruption institutions (which actually investigated all the high-level corruption cases) and help with developing processes and procedures that would minimize opportunities for corruption. After all, they should understand the integrity paradox and appreciate the enormous progress Ukraine has made in fighting corruption. 

References

1. Costa, Samia. 2013. “Do Freedom of Information Laws Decrease Corruption?” Journal of Law, Economics, and Organization 29(6): 1317-1343.
2. Gorodnichenko, Yuriy, Ilona Sologoub, Yuriy Fegyk (2026). “The Word Is Not Enough: Testing the Effects of Information Treatments on Perceived Corruption in Ukraine”, CEPR discussion paper
3. Gutmann, Jerg, Fabio Padovano, and Stefan Voigt. 2020. “Perception vs. Experience: Explaining Differences in Corruption Measures Using Microdata,” European Journal of Political Economy 65: 101925.
4. Sarullo, Nicolas, Yuriy Gorodnichenko, Tatyana Deryugina, James Hodson, Ilona Sologoub, and Anastassia Fedyk. 2026. “Measuring Corruption from Household Income and Consumption Micro-Data: An International Perspective.” Economic Modelling 160: 107578.