In state we trust: Evidence from Poland on why we undersave for retirement

ABSTRACT

With a possible decline in public pension benefits, private savings will become critical for maintaining desired living standards, yet most people undersave. We focus on one of the possible explanations for this: trust in the state. Using data from a dedicated survey and applying structural equation modeling, this study demonstrates that Polish citizens cherish the unconditional, albeit subconscious, belief in the state as a potential rescuer of last resort in life-threatening situations. This trust plays an essential role in people’s propensity to save for the future. We discuss potential explanations of this phenomenon and its consequences for social policymaking.

KEYWORDS:

• Trust
• public pensions
• private pensions
• welfare state
• structural equation modeling (SEM)
• survey study

Acknowledgments

We thank Agata Gąsiorowska for insightful comments on the measurement scales utilized in our research. All errors are our own.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available at the Harvard Dataverse (https://doi.org/10.7910/DVN/IBLHDB).

Compliance With Ethical Standards

The research has been accepted by SWPS University of Social Sciences and Humanities Research Ethics Committee (Dec. no. 02/P/09/2019). Survey respondents have been informed about the scope of the survey, voluntary participation, and the anonymization of the collected data.

Notes

1. The first pension of those who retire in a given year over the (economy-wide) average wage at retirement (Commission 2021).

2. The social minimum is one of the poverty thresholds officially used in Poland. It defines the income level that should enable a person to reproduce life forces and maintain social bonds.

3. The selection criterion was the fulfillment of the conditions for certification (ESOMAR and PKJPA, Polish certificates of quality in questionnaire research) and ensuring adequate quality of gathering data and strict verification of this process (post hoc). The sampling frame employed in the study is based on the TERYT database, used by Statistics Poland to conduct representative surveys. It resulted in an appropriate quality of the obtained data, including its representativeness of the adult Polish population.

4. This statistic measures the proportion of variance that might be common variance. The minimum value allowing for factor extraction was set at 0.5 (Kaiser 1974).

5. In Bartlett’s test of sphericity, we verify the null hypothesis stating that there are no significant correlations between the variables. In the case of a significant test result, this hypothesis is rejected; it is recognized that there are correlations between the variables, and defining latent constructs is justified. Therefore, we expect p-values close to 0.

6. With kappa = 4 as suggested by Hendrickson and White (1964).

7. CR is a measure of internal consistency reliability.

8. This is due to the relatively small number of items in individual constructs. See Peterson (1994) for a broader discussion.

9. Values of 0.95 and higher suggest redundancy and potential reduction in construct validity (Hair et al. 2019).

10. With complex models taking into account many items and latent variables, allowing for statistically significant correlations of some error terms is usually necessary; it is not possible to include all aspects of the analyzed relationships in the model. The results of the estimation of the model, in which all the error terms’ correlations were considered insignificantly different from zero, are (in the considered case) identical in terms of sign and statistical significance of regression weights, but they slightly differ in the values of loading. However, the model does not meet some of the quality of fit requirements.

11. McDonald and Bollen (1990) define the range (−1; 1) for both statistics but without strict justification for these values.

12. With the exception of variable Q24, the deviations were not large.

13. See Schreiber et al. (2006) for details of the cutoff criteria for fit indexes.

14. The literature gives rise to a measurement problem. Vickerstaff et al. (2012, 23) point out that even the wording of trust scales may lead to measurement errors. These scales usually use terms that are universally recognized as capturing trust so that trusting behavior gains a positive moral and social value, pushing up self-reported levels of trust.

15. In the entire third sector, public sources accounted for almost 43% of revenue in 2018 and donations and private contributions, less than 20% (Goś-Wójcicka 2020, 70).

16. Withdrawal is possible before reaching retirement age but results in loss of tax benefits.

17. These voluntary savings schemes are: Employee Pension Plans (PPE), Individual Retirement Accounts (IKE), Individual Retirement Protection Account (IKZE). Please see:

https://www.knf.gov.pl/knf/pl/komponenty/img/Oprac_IKE_IKZE_2019_67694.pdf; https://www.knf.gov.pl/knf/pl/komponenty/img/RAPORT_PPE_w_2017.pdf

18. The largest panel study on the quality of life of the Poles.

19. The only exception here is the Long-term Saving construct, for which the average variance extracted is a little below the assumed threshold of 0.5, meaning that the proposed items explain slightly below 50% of the variance of this latent variable; however, this is not a significant violation.

20. See OECD data, https://stats.oecd.org/Index.aspx?DataSetCode=EAG_EARNINGS

21. Having said that, as we wanted to make sure that our findings on the relationship between Trust in State and Pension-Saving Propensity are reliable, we did analyze separately the model with Income used as variable. The main results turned out to be robust. Some of the details are discussed in the next subsection.

22. For the subgroup of better-educated respondents, the standardized load was 0.120 (p = 0.070) and for the subgroup of other respondents: 0.003 (p = 0.947).

23. Standardized load for the whole sample is 0.446 (p = 0.000); for the less-educated and better-educated respondents, it is 0.406 (p = 0.000), and 0.540 (p = 0.000), respectively.

24. In the model, we added an observable variable “Income” and defined its relations to latent variables: Trust in State and Pension-Saving Propensity. As an income measure, we used equivalent monthly household income, applying the square root of the number of people in the household as the equivalence scale. Income ranges declared by the respondents were replaced by the arithmetic mean of their lower and upper bounds. The exception was the first class (below 2000–replaced with 2000) and the last class (above 15 000–replaced with 15 000). However, changing income definition to individual income does not alter the conclusions. Results of all these additional models are available upon request.

25. It recommended a direct guarantee in the Polish Constitution on the privacy of third-pillar assets.

Additional information

Funding

This work was supported by the National Science Centre (Poland) under Grant DEC-2017/25/B/HS4/00186.

Notes on contributors

Marek Kośny

Marek Kośny is an Associate Professor in the Faculty of Economics and Finance at Wroclaw University of Economics and Business. In his research, he concentrates on economic security and various aspects of distributional analysis. His recent publications include papers in Social Indicators Research, Post-Communist Economies, and the Research on Economic Inequality Series.

Radosław Kurach

Radosław Kurach is an Assistant Professor in the Faculty of Economics and Finance at Wroclaw University of Economics and Business. In his research, he concentrates on pension system design with a special focus on capital. His recent publications include papers in the Journal of Pension Economics and Finance, Applied Economics, and Applied Economics Letters.

Paweł Kuśmierczyk

Paweł Kuśmierczyk is an Associate Professor in the Faculty of Economics and Finance at Wroclaw University of Economics and Business. In his research, wok he concentrates on experimental economics and pension system design. His recent publications include papers in the Journal of Pension Economics and Finance, Applied Economics, and Applied Economics Letters.