News and Views on Public Finances: A Survey Experiment

How to cite

Draeger, Lena; Behringer, Jan; Dullien, Sebastian; Gechert, Sebastian. 2023. "News and Views on Public Finances: A Survey Experiment." AEA RCT Registry. AEARCTR-0012654. https://doi.org/10.1257/rct.12654-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Lena Draeger* , Jan Behringer , Sebastian Dullien , Sebastian Gechert

* Primary investigator

Registry ID
AEARCTR-0012654
DOI
https://doi.org/10.1257/rct.12654-1.0
First registered
Last updated
Status
Completed
Trial dates
2021-09-28 – 2021-10-12
Countries
Germany ()
Secondary IDs
D83, E60, H31, H60
Registry record
http://www.socialscienceregistry.org/trials/12654

Abstract

We use novel German survey data to investigate how perceptions and information about public finances influence attitudes towards public debt and fiscal rules. The study is conducted in an online survey with about 3.800 Germans, where the sample is chosen to be representative of the overall population.
Our survey experiment is conducted in two steps: In the first step, we randomly split the sample into three groups. The control group receives no anchor, whereas the debt-anchor groupand the interest-anchor group are informed about one historical value of the German debt-to-GDP or interest-to-tax-revenue ratio, respectively. In the second step of the experiment, we randomly provide information on the time series of the German debt-to-GDP or interest-to-tax-revenue ratios up to the most recently available year. Within these treatment groups, some respondents previously received the anchor, and some did not. After the information treatment, we measure respondents' views on public debt, on public investment and on the design of the debt brake.

Intervention

First step of the experiment:
T1: information on the German debt-to-GDP ratio in 2009 (73 percent)
T2: information on the German interest-to-tax-revenue ratio in 2009 (5.9 percent)
T3: no information (control group)

Second step of the experiment:
T1: Text "In 2020, the public debt ratio in Germany was 70 percent. This means that the debt level was about as large as two-thirds of annual economic output. The following chart shows the debt ratio in Germany over time. As a reminder: You estimated the public debt ratio in 2020 at [own estimate] percent." and graph showing the time series of the German debt-to-GDP ratio from 1991-2020.
T2: Text "In 2020, the interest-to-tax revenue ratio in Germany was about 1.4 percent. This means that out of every 100 euros of revenue, the state had to spend 1.40 euros on interest payments. The following chart shows the interest-to-tax revenue ratio in Germany over time. As a reminder: You estimated the interest-to-tax revenue ratio in 2020 at [own estimate] percent." and graph showing the time series of the German interest-to-tax revenue ratio from 1991-2020.
T3: no information treatment (control group)

Experimental design

Our survey experiment is conducted in two steps: In the first step, we randomly split the sample into three groups. The control group receives no anchor, the debt-anchor group is informed about the debt-to-GDP ratio in 2009 (73 percent) and the interest-anchor group is informed about the interest-to-tax-revenue ratio in 2009 (5.9 percent). We then evaluate the impact of these anchors on respondents' estimates of the current debt-to-GDP as well as the current interest-to-tax-revenue ratio for interests paid on government debt in Germany.
In the second step of the experiment, directly after estimating the current ratios, the sample is randomly split again: the first group receives the debt info treatment. Of these, 50 percent were previously treated with the debt anchor. The debt info treatment provides text and graphical information about the time series of the debt-to-GDP ratio in Germany and reminds respondents about their previous estimate of the current debt-to-GDP ratio. The second group receives the interest information treatment. Once more, 50 percent of respondents in this treatment group had been given an interest coverage anchor prior to the information treatment. The interest information treatment provides text and graphical information about the time series of the interest-to-tax-revenue ratio in Germany and reminds respondents about their previous estimate of the current interest-to-tax-revenue. Finally, the third group serves as a control group and receives no information treatment. This group also received no anchor in the first step.

Design details

Randomization method
randomization done by computed random variable during the online survey
Randomization unit
first step: randomization at the individual level second step: randomization at the group level: - T1: 50% from T1 in step 1, 50% from T3 in step 1 - T2: 50% from T2 in step 1, 50% from T3 in step 1 - T3: 100% from T3 in step 1
Clusters
4,700 individuals
Observations
4,700 individuals
Arms
Step 1: T1: 600 individuals T2: 600 individuals T3: 3,500 individuals Step 2: T1: 1,200 individuals T2: 1,200 individuals T3: 2,300 individuals
Primary outcomes
- estimates of current debt-to-GDP and interest-to-tax revenue ratios (in percent) - How far the current level of public debt in Germany is seen as a big problem (qualitative scale) - What respondents think the current constitutional debt brake prescribes (hence their knowledge about existing rules) (dummy variable for choosing the correct answer option) - How respondents think the debt brake should be constructed (after telling them about the current rule) (dummy variables for each suggested debt brake) - How respondents think public investment should be financed (dummy variables for each financing option)

Cited by

No citation data available for this registration yet.

Keywords