Preferences for Pension System Size Along the Bismarckian-Beveridgean Spectrum

How to cite

Anguita, Ernesto; Vujić, Sunčica; Deschacht, Nick. 2025. "Preferences for Pension System Size Along the Bismarckian-Beveridgean Spectrum." AEA RCT Registry. AEARCTR-0016222. https://doi.org/10.1257/rct.16222-1.1

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

Ernesto Anguita* , Sunčica Vujić , Nick Deschacht

* Primary investigator

Registry ID
AEARCTR-0016222
DOI
https://doi.org/10.1257/rct.16222-1.1
First registered
Last updated
Status
In development
Trial dates
2025-09-28 – 2025-10-31
Countries
Private
JEL codes
H23, H24, H55, J38
Registry record
http://www.socialscienceregistry.org/trials/16222

Abstract

Cross-national evidence shows that pension schemes with more equal benefit distributions (Beveridgean schemes) tend to be smaller than those that closely mirror individuals’ prior earnings or contributions (Bismarckian schemes). However, empirical evidence on the microfoundations of this relationship is missing. This gap limits the ability to infer the social limits of the redistributive power that can be attained through public pensions and to understand the relative importance of economic and social motivations for individuals’ preferences. To fill this gap, this experiment investigates the effect of different levels of pension scheme's transfer progressivity on individuals' preferences for the size of the scheme, measured by the contribution rate they are willing to apply.

Intervention

Individuals are randomly assigned to pension schemes that differ in their levels of transfer progressivity. We use the global measure of intragenerational redistribution proposed by Klos et al. (2022) to simulate 5 State Pension designs. Individuals are asked to state the employee contribution that they consider appropriate for the pension scheme to which they are assigned.

Additionally, we implement treatments that vary the level of government subsidy to the pension scheme and the method used to calculate individuals’ lifetime earnings. While the former affects both individuals’ own pension benefits and those of the highest and lowest contributors, the latter affects only individuals’ own pension entitlements.

Pension entitlements are determined by employees' contribution rate to the pension scheme. Respondents are able to dynamically observe how a pension design scenario with a specific contribution rate level affect 1) their current earnings, 2) their future pension outcomes, 3) the future pension outcome of top contributors, and 4) the future pension outcomes of lowest contributors.

References:
Klos, J., Krieger, T., & Stöwhase, S. (2022). Measuring intra-generational redistribution in PAYG pension schemes. Public Choice, 190(1), 53-73.

Experimental design

We recruit 1,500 working age individuals residing in Great Britain using Prolific. The experiment has been desgined in Qualtrics.

The experiment comprises three main parts that are shown to participants once they have given consent to participate.

First, they are presented with a questionnaire about their current and past labour market experience. This information is used to project each individual’s lifetime earnings and pension entitlements under different levels of transfer progressivity and contribution rates.

In the second part, participants are informed of their position on the lifetime earnings distribution and are presented with two different fictitious pension schemes. These schemes differ in how pension entitlements are distributed. Participants are able to see the impact of different contribution rates on their current net earnings and their future net pension, as well as on the minimum and the maximum net pensions available across the population. Additionally, they observe the share of the population that will receive the minimum and the maximum pension.

The experiment ends with a second questionnaire asking about their financial expectations, their financial profile and their social perceptions.

Design details

Randomization method
Qualtrics randomizes the designed treatments among participants.
Randomization unit
Randomization is made at the individual level.
Clusters
Our observations are not clustered.
Observations
1,500 individuals.
Arms
There are three treatment levels distributed across respondents. Participants are first randomized on the level of transfer progressivity that they receive in the first scheme they receive. 25% of participants (~375) receive a strictly Bismarckian pension scheme. 25% of participants (~375) receive a near Beveridgean pension scheme. 16.67% of participants (~250) receive a Beveridge-leaning encompassing scheme. 16.67% of participants (~250) receive a Bismarck-leaning encompassing scheme. 16.67% of participants (~250) receive a Bismarckian scheme. Participants are also randomized on the degree to which the pension scheme is subsidized by the government. 33.33% of participants (~500) receive a pension scheme with a low government contribution. 33.33% of participants (~500) receive a pension scheme with a medium government contribution. 33.33% of participants (~500) receive a pension scheme with a high government contribution. Finally, participants are also randomized on the way in which their lifetime earnings are calculated. 50% of participants (~750) are assumed to earn 100% of the earnings they would receive if they followed the path of a synthetic worker in the same sector and remaining in the same earnings decile within that sector throughout their lifetime. 50% of participants (~750) are assumed to earn 90% of the earnings they would receive if they followed the path of a synthetic worker in the same sector and remaining in the same earnings decile within that sector throughout their lifetime.
Primary outcomes
Preferred employees' contribution rate. This variable is used as a measure of the preferred size of the pension scheme given a certain level of transfer progressivity.
Secondary outcomes
Preferred pension scheme. This variable will measure whether respondent prefer the more progressive pension scheme, the less progressive one or none of them.

Cited by

No citation data available for this registration yet.

Keywords