x

Please fill out this short user survey of only 3 questions in order to help us improve the site. We appreciate your feedback!
Inequality, Information, and European Income Tax Policy Preferences
Last registered on March 23, 2019

Pre-Trial

Trial Information
General Information
Title
Inequality, Information, and European Income Tax Policy Preferences
RCT ID
AEARCTR-0004044
Initial registration date
March 22, 2019
Last updated
March 23, 2019 8:28 PM EDT
Location(s)
Region
Primary Investigator
Affiliation
University of North Carolina at Chapel Hill
Other Primary Investigator(s)
PI Affiliation
NYU, Abu Dhabi
PI Affiliation
Stanford University
PI Affiliation
UCLA
Additional Trial Information
Status
In development
Start date
2019-03-23
End date
2019-05-01
Secondary IDs
Abstract
It is well known that economic inequality has risen in many democratic states over the last several decades; yet there is little evidence that these states have responded with tax and transfer policies that would significantly mitigate the impact of these changes on post-tax and transfer outcomes (e.g., Bonica et al. 2013). The lack of a policy response to rising inequality has two broad interpretations. The first is that democracy is failing to deliver the policies that voters want, either because of representation failures (due, e.g., to money in politics) or because states are constrained by efficiency considerations from meeting voter demands (Gilens 2012). The second is that voters do not necessarily want rising inequality to be corrected by more generous tax and transfer policies (Ballard-Rosa, Martin & Scheve 2017). One obviously important way to assess the relative strengths of these two interpretations is to determine what the policy preferences of voters are and whether they are influenced by information about inequality. Our proposal focuses on these questions.

The tax system is one of the most powerful tools with which states can influence inequality. A key characteristic of modern tax systems is that they are multidimensional -- countries can tax different types of economic activities, including income, wealth, property, and consumption; and they often choose to tax different levels of each of these at different rates. This creates a challenge for characterizing the policy preferences of voters over the tax system and in particular its features, such as progressivity, that influence its impact on inequality.

In this research, we propose fielding a conjoint survey experiment on a representative sample of adults in Germany to study tax rate preferences across the income distribution while taking account of the impact that alternative tax plans have on total revenue raised. Our methodology not only allows us to characterize how popular alternative rates are but provides evidence of the elasticities of support for income tax plans with varying rates across the income distribution, and for taxes on income versus ones on consumption (e.g., VAT). These elasticities of public support have wide-ranging implications for the political feasibility of varying tax policy reforms.

In addition to employing an innovative methodology to measure multi-dimensional tax preferences, the research proposed here will also investigate the impact of inequality on tax policy preferences. Prior research on the question of how, or even whether, inequality affects redistributive preferences, including tax policy, is primarily observational. Researchers have investigated cross-sectional correlations of support for redistribution, or for taxing high incomes, with national or sub-national inequality, and some have traced such correlations over time (Alt & Iversen 2017, Kelly & Enns 2010, Lupu & Pontusson 2011, Moene & Wallerstein 2001). While informative, these studies are often difficult to interpret because levels, and even changes, in inequality (and certainly perceptions of either) are endogenous to policy and policy opinions.
External Link(s)
Registration Citation
Citation
Ballard-Rosa, Cameron et al. 2019. "Inequality, Information, and European Income Tax Policy Preferences." AEA RCT Registry. March 23. https://doi.org/10.1257/rct.4044-1.0.
Former Citation
Ballard-Rosa, Cameron et al. 2019. "Inequality, Information, and European Income Tax Policy Preferences." AEA RCT Registry. March 23. http://www.socialscienceregistry.org/trials/4044/history/44015.
Sponsors & Partners

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

Request Information
Experimental Details
Interventions
Intervention(s)
Voluntary adults (18+) in Germany recruited by YouGov will be contacted and asked to complete an online survey from their computer or cell phone. The survey will include a number of informational interventions dealing with the extent of inequality in society, as detailed below.
Intervention Start Date
2019-03-23
Intervention End Date
2019-05-01
Primary Outcomes
Primary Outcomes (end points)
Our key outcome of interest in this study is the multidimensional set of preferences for tax rates, as identified via our conjoint experiment (described in detail in the supplementary document detailing the full survey text). A second measurement strategy for assessing preferences over tax progressivity will elicit rate preferences directly across the set of current tax brackets. In essence, we are interested in identifying the degree of progressivity or regressivity in preferred tax systems in the German public, following the methodology developed in Ballard-Rosa, Martin & Scheve (2017), and the sensitivity of these preferences to information about inequality.
Primary Outcomes (explanation)
In our conjoint experiment, respondents will be presented with pairs of hypothetical tax plans. These plans will vary by the marginal tax rate imposed on each income bracket (as determined by actual current tax policy), as well as the Value-Added Tax rate (VAT). Each of these tax rates will be drawn randomly from an underlying set of possible values. Respondents will then be asked, in each instance, whether they prefer plan A or plan B. This binary choice outcome will be used to construct, at the level of an individual tax plan, our main outcome of Tax Plan Favored.

In addition to the forced pairwise comparisons that will be used to generate our primary outcome of interest, respondents will also be asked to evaluate each plan separately on an 11 point scale (from 0-10), with a zero representing “strongly oppose” and a 10 representing “strongly support.” These scalar measures will be used to create, at the level of an individual tax plan, our secondary measure of Tax Plan Support.

Finally, we will ask respondents to select their ideal tax rates across existing German income tax brackets.
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
While our primary outcomes of interest relate to individual support for a given tax plan (as described immediately above), we will additionally construct measures to evaluate the proposed mechanisms highlighted by our theory. To assess both the viability of our experimental intervention, as well as the causal path via which information about inequality may affect tax policy preferences, we will additionally construct two mediation checks. The first, following results from the US in Kuziemko et al. (2015), will ask respondents whether they feel that “economic inequality is a serious problem in Germany,” with responses ranging from “Not a problem at all” to “A very serious problem.” As a manipulation check, we will regress this measure of Concern about inequality on our informational treatment dummies, in order to assess whether this information did, in fact, affect views over the seriousness of inequality as an issue.

We have also included a series of manipulation checks assessing respondent information about inequality. This will involve questions asking respondents to guess where they fall in the overall income distribution (by asking them to guess “what percent of Germans with incomes would you guess earn more than you”), as well as questions regarding income thresholds for the top 1%/median/bottom 10% (“The X% income threshold is the minimum annual individual income above which a person officially earns more than [100-X]% of Germans with incomes. Above what annual income before taxes would you guess that a person earns more than [100-X]% of Germans with incomes?”). These measures will serve two functions: first, we will regress a constructed outcome measure Correct response for each of these informational questions (based on whether respondents provided the correct answer for a given threshold) on our treatment dummies to assess whether respondents that were provided with this information actually are more likely to correctly answer these questions (indicating that respondents did, in fact, update their priors about the income distribution). Secondly, we will use responses to these questions from the control group (which will not receive any information) as a measure of the average views on inequality in society.
Experimental Design
Experimental Design
Beyond the conjoint experiment described above, respondents will also be exposed to a set of “informational treatments” regarding the extent of inequality in their society. These informational treatments build off prior work by Kuziemko et al. (2015), which has demonstrated the effects of information on inequality on American respondents’ political preferences.

More precisely, each respondent will be randomly exposed to one of four informational treatments, or instead will receive no additional information (control). Each of the informational treatments is detailed immediately below:

Module A (“Personal inequality - Earn less”):
Respondents will first be asked to enter their “gross monthly individual income before taxes.” After entering this information, the survey will display the following text: “X% of German individuals who have incomes earn less than you.” In this instance, X% will be populated based on data derived from the LIS on the current distribution of income in Germany.

Module B (“Personal inequality - Earn more”):
Respondents will first be asked to enter their “gross monthly individual income before taxes.” After entering this information, the survey will display the following text: “X% of German individuals who have incomes earn more than you.” In this instance, X% will be populated based on data derived from the LIS on the current distribution of income in Germany.

Module C (“Societal inequality – Top/Middle”):
Respondents will be presented with the following statement: “In 2015, German individuals with incomes in the top 1% earned at least 124,997 euros annually before taxes while German individuals with incomes in the middle of the income distribution (where half of individuals earned more, and half less) earned 21,000 euros annually. This means that in 2015 the ratio between the top 1% of earners and those at exactly the middle was 5.95.”

Module D (“Societal inequality – Top/Bottom”):
Respondents will be presented with the following statement: “In 2015, German individuals with incomes in the top 1% earned at least 124,997 euros annually before taxes while German individuals with incomes near the bottom of the income distribution (bottom 10%) earned at most 4,536 euros annually. This means that in 2015 the ratio between the top 1% of earners and the bottom 10% was 27.56.”

Control Module
Respondents assigned to the control condition will not be presented with any additional information prior to moving to the next section of the survey.

To help us interpret the mechanisms driving the expected treatment effects in the analysis, we plan an eliminated effect experiment at the end of our survey. Specifically, we seek to determine if the impact of information about inequality has its impact because respondents have an ability-to-pay fairness criterion in which it is viewed as fair to tax the rich at a higher rate because they are better ability to afford it.

The respondents will be randomly assigned into one of the four following conditions:

CONDITION A
[CONTROL, NO INTRODUCTION]

CONDITION B
In 2015, 0.01% of Germans who earned income earned one million euros or more. This income is 48 times the median German income of 21,000 euros per year.

CONDITION C
Some argue that tax systems should have higher rates for those with higher incomes becomes high earners have a greater ability to pay and therefore the higher rates equalize sacrifice in the tax system.

CONDITION D
In 2015, 0.01% of Germans who earned income earned one million euros or more. This income is 48 times the median German income of 21,000 euros per year. Some argue that tax systems should have higher rates for those with higher incomes becomes high earners have a greater ability to pay and therefore the higher rates equalize sacrifice in the tax system.

They then will be asked a new outcome question eliciting their preferred tax rate for respondents who make more than one million euros per year.

We will then estimate the eliminated effect as in (Acharya, Blackwell, and Sen 2018).
Experimental Design Details
Randomization Method
Computer randomization
Randomization Unit
Individual respondent
Was the treatment clustered?
No
Experiment Characteristics
Sample size: planned number of clusters
N/A
Sample size: planned number of observations
2,000 individuals, each asked to make 8 choices (and therefore evaluating 16 plans) for a total of 32,000 tax plan observations
Sample size (or number of clusters) by treatment arms
400 respondents (evaluating 4800 tax plans) in Module A, 400 respondents (evaluating 4800 tax plans) in Module B, 400 respondents (evaluating 4800 tax plans) in Module C, 400 respondents (evaluating 4800 tax plans) in Module D, 400 respondents (evaluating 4800 tax plans) in Control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

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

Request Information
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
NYUAD Institutional Review Board
IRB Approval Date
2018-08-15
IRB Approval Number
046-2017
IRB Name
Administrative Panel on Human Subjects in Non-Medical Research
IRB Approval Date
2018-05-07
IRB Approval Number
39652
IRB Name
Office of Human Research Ethics
IRB Approval Date
2018-05-17
IRB Approval Number
16-2946
Analysis Plan

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

Request Information
Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
No
Is data collection complete?
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)
REPORTS & OTHER MATERIALS