Reducing the Shadow Economy with Nudges: A Randomized Control Trial in Latvia

Last registered on April 28, 2023


Trial Information

General Information

Reducing the Shadow Economy with Nudges: A Randomized Control Trial in Latvia
Initial registration date
September 07, 2021

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
September 09, 2021, 7:15 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
April 28, 2023, 6:16 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.



Primary Investigator

University of Latvia

Other Primary Investigator(s)

PI Affiliation
University of Bonn, Institute for Applied Microeconomics

Additional Trial Information

Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
The goal of the study is to investigate the behaviour of corporate firms in Latvia and examine various approaches to increase the tax compliance. The interventions of the study examines two competing theoretical explanations what motivates to pay taxes. One is the so-called deterrence approach, stemming from the standard model of tax evasion. It predicts that tax compliance is based on tax rate, audit probability, penalty rate and income. Alternatively, non-deterrence approach states that tax compliance is based on broader propositions, such as morality, social norms and willingness to contribute to the public good. The field experiment examines both approaches on tax compliance behavior in Latvia. The study consists of seven different messages randomly assigned and sent to firms for whom the average salary in the first half of 2021 has been lower than 70% in comparison to the region and sector in which the firm is operating.
External Link(s)

Registration Citation

Chapkovski, Phillip and Andris Saulitis. 2023. "Reducing the Shadow Economy with Nudges: A Randomized Control Trial in Latvia." AEA RCT Registry. April 28.
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details


The control message asks to minimize the risks of engaging in the shadow economy by considering to raise the salaries. Three treatment messages additionally include various audit probabilities, while the remaining three treatment messages include various non-deterrent messages.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Average salary for the employees reported to the State Revenue Service during the period October-December 2021.
Primary Outcomes (explanation)
In order to compare the deterrence to non-deterrence approach, I will merge the first three treatment groups (where the text messages state various audit probabilities) and the last three treatments (where the text messages include various non-deterrence messages) and compare them to the control group and no message condition.

Secondary Outcomes

Secondary Outcomes (end points)
(i) the firm's average salary ratio to the average salary in the region and sector the firm operates; (ii) number of employees declared; (iii) number of incoming requests from the firms receiving the text messages; (iv) number of closed down businesses.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experimental design consist of seven different text messages communicated to the firms with average salary in the first half of 2021 below the 70% average in the region and sector the firm is operating. The messages are advising to mitigate the differences in the next months. The treatment messages are examining the best approach to communicate with the firms. They are created based on previous studies on tax compliance and compares deterrence and non-deterrence approaches.
Experimental Design Details
The data collection started on July 2021 when reports on the reported salaries for the first half of 2021 were retreived. The firms with average salary less than 70% of the average salary in the region and sector (as reported by the Central Statistics Bureau of Latvia) were included in the study.

Before the intervention, we did a pilot with all seven treatment messages with 140 firms in total. The goal of this pilot was to identify the level of feedback of the firms to calculate the administrative burden the State Revenue Service will have to manage, as well as to make sure the text messages are clear and does not raise any misunderstandings, i.e., the letter is suggestive and informative only.

The treatment messages were sent in three waves during September 2021. Each firm received one message once via the Electronic Declaration System hosted by the State Revenue Service.
Randomization Method
Randomization was carried out in office by a computer. The firms were blocked by the statistical region and balanced on key control variables, namely: (i) average salary ratio to the region and sector average salary for the first half of 2021 and July 2021; (ii) the total amount of paid taxes in 2021; (iii) business operating sector; (iv) total turnout in 2021 (converted to log); (v) average number of employees in 2021; (vi) age of the firm; (vii) whether the firm has received COVID-19 crisis related grants from the state.
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
3,929 firms
Sample size: planned number of observations
3,929 firms
Sample size (or number of clusters) by treatment arms
491 firms no message, 489 control, 490 Treatment group 1, 491 Treatment group 2, 494 Treatment group 3, 490 Treatment group 4, 491 Treatment group 5, 493 Treatment group 6. The actual number of firms in each experimental condition can be different if the firm has been audited during the period July-September 2021, i.e., before the beginning of the intervention. In that case, the firm is dropped from the sample.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Intervention Completion Date
September 21, 2021, 12:00 +00:00
Data Collection Complete
Data Collection Completion Date
February 09, 2022, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

This study employs a field experiment and qualitative content analysis to examine the effect of various behaviourally-informed messages on increasing tax compliance in Latvia. In a field experiment, more than 3,000 businesses received a message with a normative appeal to increase the relatively low salaries compared to firms operating in the same industry and region. Other treatment groups received the same message with an additional paragraph that varied audit probabilities or included prosocial messages. All treatments effectively increased the average declared salaries in the enterprises relative to not sending a message. Even though the overall fiscal effect was positive, the qualitative analysis of the feedback by the firms indicates that messages, particularly those that did not state the future actions of the tax administration, provoked discontent and distrust between the taxpayer and the tax administration. Our findings demonstrate that clear communication of the intended actions of the tax administration is the most effective approach to promoting tax compliance. Furthermore, our research indicates that a relatively small audit probability (5%) is as effective as a larger probability (66%), implying that there is no need to carry out audits on a large scale to address tax evasion.
Saulitis, Andris and Chapkovski, Philipp, Investigating Tax Compliance with Mixed-Methods Approach: The Effect of Normative Appeals Among the Firms in Latvia (March 1, 2023). Available at SSRN: or

Reports & Other Materials