Mechanisms Behind Responses to Pension Projections

2026-07-01

2026-08-31

Expected pension-income benchmark deviation, confidence in pension expectations, perceived pension adequacy, zero-sum thinking, retirement-preparation behavioural intentions, and perceived agency over retirement outcomes.

Expected pension-income benchmark deviation is captured by the absolute deviation between respondents' post-treatment expected total monthly old-age pension and the relevant income-group benchmark. Confidence in pension expectations is captured by respondents' self-reported confidence in their estimate of their own future old-age pension and, where applicable, by their perceived uncertainty about future retirement income.

Perceived pension adequacy is captured by respondents' assessment of whether their expected future pension income will be sufficient to maintain a decent standard of living in retirement. Zero-sum thinking is captured by an index based on respondents' agreement with survey items measuring whether they view economic gains for one group as implying losses for others. The index will be constructed from the relevant post-treatment items, with higher values indicating stronger zero-sum beliefs.

Retirement-preparation behavioural intentions are captured by respondents' stated likelihood of taking retirement-preparation actions, including starting or increasing voluntary retirement saving, seeking pension-related financial advice, checking information about their second-pillar fund allocation, and adjusting expected retirement timing. Perceived agency over retirement outcomes is captured by the difference between respondents' perceived own influence over their financial situation in old age and their perceived dependence on state decisions, with higher values indicating greater perceived personal agency relative to dependence on the state.

Pension-system attitudes, demand for pension information, willingness to pay for personalized pension information, information avoidance, views on pension reform, intergenerational-fairness attitudes, intended sources of retirement income, investment-risk attitudes related to retirement saving, and exploratory beliefs about artificial intelligence, labour markets, and future pension income

Secondary outcomes include broader pension-system attitudes, demand for pension information, willingness to pay for personalized pension information, information avoidance, views on pension reform, intergenerational-fairness attitudes, intended sources of retirement income, investment-risk attitudes related to retirement saving, and exploratory beliefs about artificial intelligence, labour markets, and future pension income.

We will also analyze treatment-specific manipulation-check and mechanism outcomes. These include whether the information shown to respondents was perceived as surprising relative to prior expectations, whether respondents felt that the information was personally relevant to their own situation, and whether the information made them think about their everyday financial situation in retirement. These outcomes will be used to assess whether the treatments were understood, perceived as relevant, and activated the intended informational or salience channels.

Secondary confirmatory analyses will compare selected treatment arms with each other within the main outcome families. These analyses include pairwise and pooled contrasts across level-information, scenario-uncertainty, pillar-decomposition, consumption-salience, and institutional-uncertainty treatments. The pre-specified secondary contrasts compare:
(i) income-level information and scenario-uncertainty information;
(ii) pooled information provision and pension-pillar reframing;
(iii) economic uncertainty and institutional uncertainty;
(iv) consumption salience and institutional uncertainty;
(v) pooled information treatments and pooled mechanism-oriented treatments; and
(vi) pension-pillar reframing and pooled information treatments.

The study is a randomized survey experiment among Slovak residents aged 18–65 who are not currently receiving an old-age or early old-age pension. The main survey targets 6,000 respondents and uses a mixed 80% CAWI and 20% CAPI design with quotas on sex, age, education, settlement size, and region to ensure national representativeness.
The survey first measures respondents’ receipt and understanding of the official 2026 pension forecast letter, pension-system knowledge, financial preparedness, baseline pension expectations, and zero-sum beliefs. Respondents are then randomly assigned to one of six groups: one control group and five information-treatment groups. The control group receives neutral information about the Slovak pension system. The treatment groups are: (1) information on pension benefits by lifetime-income group; (2) pension projections under different economic scenarios; (3) the role of the first and second pension pillars; (4) the consumption implications of expected pension income; and (5) uncertainty in public versus private retirement saving.
After the treatment, we re-elicitate key beliefs and measure treatment reactions, behavioural intentions, attitudes toward pension reforms, and views on intergenerational fairness. Self-reported receipt of the official forecast letter is observed but not randomized and will therefore be analysed as an observational layer.

Not available

Eligible respondents will be randomly assigned to the control group or one of the five treatment groups by the professional survey agency. Assignment will be computer-generated within the survey platform.

Individual respondent.

No

Not applicable; randomization is conducted at the individual respondent level.

6,000 respondents.

Approximately 1,000 respondents per arm across six experimental arms.