Staying physically fit and economically active: recipe for a high-quality life and well-being after retirement in the Czech Republic Open Access

The increasing life expectancy of the citizens compels global society to recognize the importance of understanding the consequences of demographic changes and population ageing, which brings several challenges for policymakers and scholars. Notably, these challenges include the ongoing revision of state pension policies (Pilipiec et al., 2021; Weber and Loichinger, 2022; Tejativaddhana et al., 2022) and the adoption of the United Nations Sustainable Development Goals (SDGs), particularly in ensuring healthy lives (World Health Organization, 2024) and promoting well-being — defined as overall life satisfaction or happiness (Veenhoven, 2012; Weimann et al., 2015; Saner et al., 2020) for individuals of all ages, as outlined in the objectives of SDG 3 (United Nations, 2024).

Achieving balanced physical and mental well-being, especially in the post-retirement age, has been the focus of numerous studies, including the works of Horner (2014), Wahrendorf et al. (2017) or Hansson et al. (2020a), trying to understand the factors, characteristics and circumstances that shape the overall life satisfaction of the elderly population. The existing research on individual well-being is extensive, encompassing various empirical approaches, conceptual studies and heterogeneous observations. To illustrate this variety and breadth, the author references a systematic literature review by Diener et al. (2018), which summarized about 170,000 scientific articles, including those focusing on the well-being of the elderly population and its drivers. Diener et al. (2018) divided these drivers into several categories, including demographic factors, external circumstances, and internal characteristics. Diener et al. (2018) reported relatively small effects of demographic characteristics, such as gender or educational level, while highlighting a general trend of decreasing well-being in older age as one approaches the end of life (Gerstorf et al., 2008).

The most impactful well-being drivers, referred to as resources (Kalimo et al., 2002), are described in resource-related theoretical approaches towards ageing, such as the resource theory of retirement, or the resource-based dynamic model for retirement adjustment (Noone et al., 2022). These theories emphasize the need to plan retirement systematically, taking into account the expected decline in earnings and income. Earlier research summarized by Diener et al. (2018) concluded that among the most significant factors are income and household economics-related variables, such as income, job attitudes, and economic activity in post-retirement age, all of which are, of course, contingent upon health status, particularly physical capacities, i.e. the strength, the absence of diseases or injuries. According to Noone et al. (2022), financial planning or combining retirement with work in post-retirement age might assist in managing time and mitigating the financial gap that results from leaving an economically active career. The lay theory (Bonsang and Klein, 2012) suggests that having more financial resources might help individuals buy even higher levels of well-being. However, current research (Hansson et al., 2020a; Soepding et al., 2021; Noone et al., 2022) concludes that the evidence is not fully conclusive, highlighting the need for more empirical findings to document the relationship between job attitudes, income and well-being in post-retirement age. At the same time, it is important to acknowledge that well-being in post-retirement age is also influenced by household structure, family ties and social interactions, including community involvement and religious attitudes, all of which impact day-to-day life in retirement (Kim and Jung, 2021).

This research aims to contribute to the extensive body of knowledge on factors shaping well-being in post-retirement age by providing insights from the Czech Republic, a small open economy with a stable healthcare system and a pay-as-you-go pension system for retirement at age 65 (Andel, 2014; Mudrak et al., 2016; Fialova, 2018; Łuczak, 2018; Zemancová et al., 2024). The current body of knowledge on well-being drivers will be expanded by presenting findings on the interaction between economic activity in post-retirement age and health condition, measured both subjectively and more objectively through the values of Body Mass Index (BMI). This approach helps to address the gap in the prior research that has primarily relied on individual perceptions of the respondents (Diener et al., 2018; Wang et al., 2023). We utilize data from the Czech sample (N=8,435) of the annual Survey of Health, Ageing and Retirement in Europe (SHARE). In particular, we analyse survey data from 2010 to 2020 and employ multivariate regression analysis to reach the conclusion. The next chapter will further introduce the dataset, the variables used, and their descriptive findings. Then, the results from the hierarchical regression modelling will be presented and discussed in the final section, along with recommendations for future research.

This quantitative research utilizes extensive data from the annual Survey of Health, Ageing and Retirement in Europe (SHARE), a widely recognized survey for studying the economic behaviour of European citizens throughout the course of their lives and beyond. The SHARE survey is well-known within the scholarly community; therefore, the author references methodological publications that describe the data collection procedures, processes and data harmonization adjustments (Bergmann et al., 2022; Börsch-Supan, 2022; Börsch-Supan et al., 2013). The author worked with a pooled sample of the Czech population surveyed between 2011 and 2020, focusing specifically on the post-retired population in the Czech Republic age 65 and older (for details, see the European Commission (2024), which describes retirement conditions in the country, including early retirement options). Consequently, there is a sample size of 8,435 individuals with complete data records available for empirical analysis.

The key variable utilized from the SHARE sample provides information about the quality of life and well-being, measured via the CASP index, an established tool in the scholarly community measuring quality of life in later life (Sim et al., 2011; Howel, 2012). This serves as the explained or dependent variable in this study. The CASP index is a complex index that combines several components: Control, Autonomy, Self-realization and Pleasure (Sim et al., 2011; Horner, 2014), representing a relevant subjective measure of well-being. The index can reach values ranging from 0 to 36, with higher values indicating the greater the well-being.

This study investigates whether economic activity in post-retirement age is associated with higher well-being, providing the elderly population with opportunities for self-realization, utilization of the accumulated work experience, expanding financial income and a chance to interact with other people socially (Viljamaa et al., 2022; Sousa et al., 2023). The author measured this by asking the question about the employment status of the respondents, i.e., employed or self-employed. If they answered yes, either employed or self-employed, then they are considered as ongoing participation in the labour market. Unfortunately, there were no specific details of their occupation, which is a commonly known limitation of using the SHARE data.

Descriptive statistics of the variables was shown in Table 1. The dataset consists of individuals aged, on average, 73.4 years, with the eldest person at the age of 104 years. The data show that the sample is more biased towards females (66.6 percent), which could also be explained by the longer life expectancy of the women in the country (Langhamrová et al., 2012; Arltová et al., 2013; Sunwoo, 2020). When it comes to economic activity, 7.6 percent of the Czech respondents are employed or self-employed, and close to half of them (48.4 percent) live in their household together with a partner. The average household size is 1.8, which is quite common in this lifetime when kids have already grown up and lived together with their own immediate families (Vrabcová et al., 2017).

Health and physical condition are undoubtedly essential factors shaping the ability to remain economically active even in post-retirement age (Dave et al., 2008; Weber and Loichinger, 2022; Garrouste and Perdrix, 2022). This is why this study accounts for self-perceived health (scale item from 1= poor to 5= excellent) and more objective Body Mass Index (BMI), which we further classify into underweight, normal, overweight and obese. Furthermore, it takes into consideration the situation of the households, such as the number of persons living together, the household capacity to make ends meet (scale item from 1= with a great difficulty to 4= easily) and the zero-one dummy variable reflecting whether the partners live together.

The empirical analysis relies on a quantitative approach, employing hierarchical regression analysis, to examine factors shaping the well-being of the Czech elderly population. In particular, the analysis is to investigate the role of economic activity in the post-retirement age by also controlling for the household situation and health status. The hierarchical approach is common in this type of study, relying on the SHARE data, such as research by Sunwoo (2020) or studies by Wahrendorf et al. (2017) and those reviewed in the systematic literature review article by Pilipiec et al. (2021).

Table 2 presents the final estimates of the hierarchical regression analysis approach. In total, three estimated models (Models 1, 2 and 3) that differ in their specifications are reported in Table 2. They are controlled for the year of the survey, and the coefficients were estimated by the ordinary least squares (OLS) technique with robust standard errors. The Chi-square test for the joint model significance shows that each of the models is statistically significant.

The overall findings are summarized as follows. First, it is not surprising that overall well-being decreases with age. Second, there are no differences concerning gender despite the expected differences in life expectancy. On the other hand, better-perceived health is positively related to the overall CASP index. While the other proxy for physical condition and health, i.e., BMI, itself is not statistically significant in Model 1, recoded into the groups in Models 2 and 3 provides more reasonable findings, showing that higher values of BMI seem to be positively associated with overall well-being. Economic activity, i.e., being employed or self-employed, also positively influences well-being. Therefore, it was decided to explore the effect further by combining the normal BMI values, which should represent sufficient eligibility for work-related tasks, with the economic activity in Model 3. The findings support the assumption that those with normal BMI and economically active values have an additional boost in well-being values, which aligns with this study's assumptions. Furthermore, it was noted that from the household-related characteristics, the only significant factor is the financial/economic sustainability of the household, which is positively related to the CASP levels of the respondents.

This research aimed to contribute to the existing knowledge on the factors impacting the well-being of individuals in the post-retirement age, i.e., above 65 years. It did so by using the Czech data from the annual Survey of Health, Ageing and Retirement in Europe (SHARE) over the years 2010-2020. The findings from the hierarchical regression modelling support the community's findings and extend those, especially within the interactions between health status and economic and income-related variables that were not fully explored by the previous studies (Diener et al., 2018; Kim and Jung, 2021; Noone et al., 2022).

In line with the conclusion of Kim and Jung (2021) and Pilipiec et al. (2021), we observe a negative impact of increasing age on overall well-being, measured in this study by the CASP index, an established construct in the scientific community (Howel, 2012; Horner, 2014). This is expected, as the increasing age is associated with even higher medical risks and worsened mental and physical conditions (Gerstorf et al., 2008). As expected from the systematic literature review of Diener et al. (2018), there was no significant difference across the genders, which are deemed to be overcome by other variables, especially health-related and economic-related variables. The author attempted to do an additional robustness check, which was limited by the significant drop in the number of observations in the sample. This additional analysis tested the effect of the highest educational attainment variable. The results, however, did not find the role of education to be statistically impactful in the SHARE data, representing the respondents from the Czech Republic. In this way, the findings align with the current research, noting that the level of education does not seem to be the influential driver of well-being for the older persons.

Conclusions from the hierarchical regression models are also straightforward in the role of the perceived health status, clearly associated with better life quality and well-being among respondents with better health. Yet as criticized by Wang et al. (2023) and Yuan et al. (2023), the author wanted to use more objective measures of physical conditions, so the Body Mass Index (BMI) was employed to extend the findings. Interestingly, the BMI effect was firstly dominated by the perceived health variable, so they were coded according to the categories used in the healthcare industry, i.e., underweight, normal, overweight and obese. The effect was tested again, and it was observed statistically for significant results. Surprisingly, the highest effect on well-being was found in the group of respondents with obese and overweight scores of BMI. The researchers dived even more into this finding and explored the linkages between BMI and economic variables in the form of statistical interactions. Nevertheless, the only significant and positive finding was the combination of normal BMI and economic activity of the respondents, i.e., being employed or self-employed. To explain this, we could use the resource-related theories of retirement (Noone et al., 2022) by assuming that respondents with the highest BMI scores, i.e., obese, do not need to be or cannot be active anymore in the labour market to reach higher values of well-being, while those with a good physical condition, could still use employment or self-employment to reach higher values of well-being, as noted by Fialova (2018) or Zemancová et al. (2024).

This theoretical approach assumes that retired individuals engaged in economic activity or with a better financial situation of the household might achieve even higher levels of well-being (Bonsang and Klein, 2012; Amani and Fussy, 2023). The analysis of the employment status partially supported this assumption, yet the lack of data on the income and wealth of the respondents limits the probability of such a causal link. Therefore, using these variables in forthcoming research could be helpful in order to test the empirical validity of the stated hypothesis. Additional variables, such as social benefits, should also be considered in future research efforts (Huxhold et al., 2014). Despite that, the findings illustrate the need to carefully plan retirement from a financial perspective (Andel, 2014; Lam, 2022; Hasegawa et al., 2020, 2023; Cassanet et al., 2023; Hasmanová Marhánková and Soares Moura, 2023) and thus provide a serious recommendation for the Czech government and non-governmental and non-profit organizations in contact with the pre-retirement population. The best practices accumulated by the scientific community highlight the support of pre-retirement clubs and gatherings, where the key questions of future retirement, including financial and economic activity, are discussed (Mudrak et al., 2016; Fialova, 2018), but of course, they cannot be used as a one-size-fits-all approach.

Yet, within the Czech social policy agenda, these actions need to be specified carefully and linked with measurable objectives, as Pospíšilová and Kalenda (2023) recently noted. Since the 1990s, there has been an ongoing policy debate on how to systematically craft ageing policies, noting the need to promote active labour market participation among the elderly population (in line with the provided findings), yet due to other political goals and lack of financial resources distributed towards shaping those policies systematically, this policy discussion remains unsolved (Perek-Białas et al., 2006; Potůček, 2023; Vostatek, 2024). This study aims one more time to attract the interest of policymakers and politicians towards the phenomenon and highlight the need to address the issue on time, as tailoring social and ageing policies requires the support of consensus among the various political parties and wings as a condition of successful implementation.

For future research, this study has highlighted that more efforts of the scientific community need to be centred on the role of labour market activity and economic sustainability of the elderly population by using more accurate measures rather than relying on individual perspective variables to advance scientific rigorousness of the current state of knowledge. In this domain, the author recommends better operationalizing variables that could strengthen our results but were unavailable in our SHARE data. These include, for example, more detailed data about the occupation of the respondents in the post-retirement age, such as type of occupation, working hours, earnings and the proportion of this income to the state pension. Such information could allow us to investigate the linkages with physical and mental conditions further and reveal in which occupations elderly individuals engage most frequently.

Another limitation of the sample is that without more control variables, made it unable (due to sample size) to work with the personality traits and characteristics, which were also found by previous studies to be impactful (Diener et al., 2018; Hansson et al., 2020b). In this manner, it is recommended that future researchers may add multiple control variables and enhance the scientific rigour of the findings through further empirical validations and tests. Lastly, despite using a cross-sectional dataset with respondents from over a ten-year period, it cannot fully refrain from the biases related to the data collection procedures and representativity of the sample to the Czech elderly population, so it is recommended to triangulate the presented findings with other datasets, such as from European social survey perspective.

Funding: This work was supported by data from SHARE Waves funded by the European Commission, DG Employment, Social Affairs and Inclusion; from the German Ministry of Education and Research; the Max Planck Society for the Advancement of Science; the US National Institute on Aging and other national funds.