RETIRE – WHY YES, WHY NOT? RESULTS OF A STUDY ON THE CZECH SENIORS

Currently, one of the most discussed topics in Europe is population aging and its consequences. Most countries have already introduced many measures to support older people and have changed their retirement age. The number of people older than 50 is increasing, and it is of great importance that this group be approached on different levels. Retirement is considered as a crucial stage of life. Therefore, the need for examination of factors influencing retirement decisions is vital. In general, there are similar factors in every country, but there are also country-specific factors, which are unique. This article focuses on the factors which are significant in the Czech Republic. The present study aims to determine the factors of retirement in the Czech 50+ population. The analysis is based on data from the SHARE database, and the factors are investigated using the logit model. The research results show that there are factors as health issues detected in other countries but also factors that are unique in the Czech Republic.


INTRODUCTION
The current retirement policy very much influences the retirement policy. Could we ask about retirement time if we had no guarantee of a pension at a certain age? The answer to such a question is by no means easy, as the pension systems vary across countries, with each country being at a different development stage. The only thing we can be confident about is that the wrong retirement age set may heavily burden the state budget and the economic and psychic ease of seniors.
In addition to hampering the state budget, retirement affects the mental health of seniors. By an EU study, retirement has a prolonged negative effect on men and women's mental health of different education levels and employment. Thus, Heller-Sahlgren (2017) suggests postponing the retirement time as much as possible, which would be beneficial to the mental health of pensioners and upholding the pension system (Heller-Sahlgren, 2017). Another benefit he sees is in the reduction of public healthcare spending. His research has been carried out using the SHARE database, which has provided data for this paper.
The present study aims to determine the factors of retirement in the Czech 50+ population. Its primary data source is the SHARE.

LITERATURE REVIEW
The decision to retire may be influenced by many factors at personal, family, organizational, and government levels. Below is a list of factors from various research projects with some of them described.
Factors influencing the retirement timing Fisher et al. (2016) are attempting to specify retirement timing (G. G. Fisher et al., 2016). To this end, they have created a model based on the currently available research projects consisting of three principal parts: previous circumstances, influencing factors, and consequences. Each of those parts breaks down into further components. The first two parts determine the time a senior leaves the labour market (prematurely, at the expected age, after the expected age), which in turn has consequences at a macro-economic level, for the family, for society, and, last but not least, on the retiring senior.

Premature retirement
Looking at the history of people retiring prematurely, we see that, several decades ago, premature retirement was receiving support from the government policy. In the 1990s, this scheme was contested mostly because of its negative influence on the pension system budgets. Still, such programs have remained common practice despite the questions raised (Courtioux et al., 2006). At present, premature retirement is seen as a threat that jeopardizes the pension system budgets.
Factors influencing premature retirement According to many research projects, a large number of factors influence the decision to retire prematurely. It is dividing such factors into two groups. Called personal factors, the first group includes factors related to the personality traits of an older person. Another group contains factors that surround a person, called environmental factors.
According to Paul and Townsend (1992), the personal factors influencing premature retirement include type-A behaviour caused by outdated competencies, health conditions, and economic welfare. Type A persons set themselves ambitious goals, work with maximum effort, hate wasting time, manage several parallel activities, are very competitive, and are in permanent stress.
Competences get outdated if a person has a feeling that his/her competences are no longer sufficient. Kaufman (1974) defines competence outdating as a stage at which specialists no longer have the current knowledge or skills necessary to maintain efficient performance while carrying out the current or future job (Hall & Kaufman, 1975). For example, Gudanowska et al. (2018) state the most popular competencies are professional knowledge and technical skills in the production industry (Gudanowska et al., 2018). According to the European Centre for the Development of Vocational Training (CEDEFOP), competence outdating may be divided into two basic types(European Centre for the Development of Vocational Training (CEDEFOP), 2010). The first type relates to physical abilities while the second one to economic skills, as competencies that used to be required for the job but are now no longer necessary.
Health condition may be regarded as a variable significantly influencing the decision to retire prematurely (Paul & Townsend, 1992). Leinonen et al. (2016) note that people who regard their health as ill are more likely to retire prematurely.
The last of the personal factors is economic welfare. According to the Italian Statistical Institute (2017), it encompasses variables measuring the economic welfare: income, wealth, expenditures, residence conditions (Economic Well-Being, 2014).
The second group of factors is those of the environment, including the implementation of work goals, the required family situation, and leisure activities.
Several research projects mention education as influencing premature retirement. In their study, The Long Reach of Education: Early Retirement (2015), the economists Venti and Wise analyze the relationship between the highest education achieved and premature retirement. By their reasoning, there are four ways in which education may indirectly influence early retirement (Venti & Wise, 2015). These include health conditions, employment, income, and accumulation of assets. However, education also has a direct effect on premature retirement. They believe that the lower the education, the more likely it is that early retirement will occur.
There are a lot of factors playing a role in retirement timing. Some factors are closely connected to a person and some to the environment where the person lives. These factors also differ from one culture to another. This paper examines the factors important in the Czech Republic.

METHODOLOGY
This paper's quantitative research used the secondary data of the Survey of Health, Ageing, and Retirement in Europe (SHARE) database. Data from different modules of the sixth, fifth, and fourth waves have been merged with each respondent's unique code. The secondary data have been analysed using descriptive statistics and logistics models.
The respondents aged under 50 and those not responding, not knowing how to respond, and refusing to answer questions concerning their present work have been removed from the data. Concerning other variables, some of them had values missing, which, however, had been taken into account in the research. The final respondent sample for the logit model contained 876 responses. The respondents included in the analysis were selected by the variable if answering that they were employees or self-employed (or employed by a family company).
Looking for early retirement was another important variable, which has been re-coded to a yes or no variable with yes coded as 1 and no as 0.
A respondent gender variable was used for group comparisons carried out in the form of non-parametric tests.

Variable descriptions
The variables used for the analysis are detailed in Table 1. They were selected based on other studies on the subject of this paper. All variables connected to the health (number of chronic diseases, self-perceived health, degree of depression, sports, or activities that are vigorous, feeling left out of things) are selected according to research done by Paul and Townsend (1992). Job-related variables (satisfaction with work, present job requires computer skills, the job involves hard work, number of hours worked per week) are selected according to Fisher et al. (2016). Variable education degree is selected according to Venti and Wise (2015) and variables household income per person and quality of life and wellbeing according to the research of the Italian Statistical Institute (2017) Table 5 Recoded Present job requires computer skills Scale: 1 "Yes"; 5 "No" it001_ Job involves hard work Scale: 1" I definitely agree"; 4 "I definitely don't agree" ep027_ Feeling left out of things Scale: 1" Often"; 4 "Never" ac016_ Self-perceived health Scale: 1 "Excellent"; 5 "Bad" Sphus Number of hours worked per week Scope ep013_ Sports or activities that are vigorous Scale: 1 "More than once a week"; 4 "Almost never" br015_ Quality of life and well-being Scale Casp Source: Author's own Education variable recoding The International Standard Classification of Education (ISCED) 97 has been recoded to match the Czech education categories (Schröder & Ganzeboom, 2014). To this end, the six levels have been reduced to three as shown in the below table in much the same way as for other studies. A total of 4858 respondents in the Czech Republic took part in the years 2014/15, including 56.39 % of women. As the presented paper emphasizes, it is substantial to know the percentages of the retired, employed, and unemployed persons regarding retirement factors. This is shown in Fig. 2: 3,737 respondents are retired, and 880 are employed. Seventy-three retired persons have multiple employments. The respondents of the selected group (N=876) are aged 50+. Their average age is 60,7 and median age 61 years, the oldest respondent being 89. Concerning the gender proportion, it is almost even with 45.4 % of men. Comparing the highest education achieved, we see that the proportions of university-educated men and women are almost equal with 19.46 % women and 20.40 % men). A difference of about 10 % can be found in secondary education. Table 3 details education in terms of gender. The overall education median is 2 (55 % of the respondents), which is secondary education. Dividing respondents into employees and those self-employed, we see that most of them are employees. These include 56.8 % of those employed in the private sector (see Figure 3). Another essential characteristic investigated by this respondent group is family or marital status. 74.4 % of the respondents are married and living in a common household. Those divorced are also strongly represented (16%). Most households consist of two persons (48.9 % of the respondents. 24.1 % of the respondents live in a three-person household (parents and child/children). A substantial variable is also the household income. The net income per household is 20130.87 CZK a month, with a median of 18981.09 CZK. Figure 4 shows the household net income. The below table displays the factors essential for the logit model. This model represents the way the desire for retirement depends on the factors. The "Look for early retirement" variable is dependent. It answers the question: "If you think about your current employment ---would you prefer retiring as soon as possible? ". In Czech, this question is not a direct question about premature retirement, but in English, it is.

Retire
As this inquiry was carried out in Czech, it will further be regarded as and called "desire for retirement. " The variables in the table are independent. The Nagelkerke R 2 for this model is 74.4 %. The limit value used for the analysis is 0.5.   Table 5 lists important logit model variables. The satisfaction with work and the number of chronic disease variables, having a negative coefficient, has a negative effect on the dependent variable. In other words, due to more chronic diseases and decreased work satisfaction, the chances of the desire for retirement variable being zero will increase, which means that the individual will desire for retirement. Although the household income variable was marked as important in the model, it has no significant effect on whether a respondent desire for retirement or not. However, the more hours worked per week, the greater the chance of less desire for retirement. If the current work requires computer skills, the chance is less of premature retirement. The respondents mostly answered the question as to the degree of depression as none or almost none. Therefore, the relationship between this variable and the explained variable is likely to be positive. Thus, the respondents with a higher degree of depression desire to stay at work (not desiring for retirement). If a respondent enjoys good health, the chance is higher if he or she does not desire to retire. The situation is much the same as the work-is-physically-demanding, feeling-left-out-of-things, and education-level variables. As acceptable are regarded variables whose p-value is about 0.1. These include sports or physically demanding activities; quality of life, and well-being.
The Exp (B) indicator or chance quotient means the change in the chance as the independent variable changes by a unit. If it is higher than one, the chance will increase.
If the value of the perceived health condition variable increases by a unit, the probability of the dependent variable changes by 13.84 % (converted by an exponential function from the chance quotient to chance and, from a chance to probability).
The factors identified by the logit model include satisfaction with work, physical (number of chronic diseases) psychological health (degree of depression, self-perceived health), the financial situation of the respondents (household per capita income), number of hours worked per week, the current job requires using a computer, job physically demanding, feeling left out of things, and level of education. The quality of life, well-being, and participation in sports or activities that are vigorous are among the less significant factors.

DISCUSSION
The results open a broad discussion on retirement timing not only in Czech society. Obviously, there are factors that significantly influence the desire for retirement. Those are satisfied with work. This influence is negative; the more the senior is satisfied with his/her work, the lower is his/her desire to retire. The next variables influencing seniors' retirement decisions have a positive effect: number of chronic diseases; degree of depression, hard work; feel left out of things; level of education.
The other tested variables were close to the level of significance and apparently, they are important for a certain group of seniors; even though they did not influence the desire for retirement significantly: household per capita income; number of hours worked per week; current work requires computer skills; health condition perceived; sports of physically demanding activities; quality of life and wellbeing.
These results are in compliance with other authors; Van den Berg (2010) identified the following factors: ill health, physically very demanding job, too much stress at work, little satisfaction with work, insufficient physical activity at leisure (Schäfer et al., 2010). A Dutch study (Van Dalen et al., 2009) adds little appreciation at work to these factors. Other studies prove that seniors' knowledge and skills and education should be among the factors (Damman et al., 2011;Herrbach et al., 2009). Some studies emphasize the social and financial factors influencing seniors (Fisher, 2006). Still, these results prove that the group of seniors is diverse, and retirement reasons vary. These findings deny the popular approach "one fits all" in setting the retirement age across countries. The portfolio of different options that would consider the above-tested variables that influence the decision for retirement timing might be a solution.
Also, authors use more often rather the healthy life years and expectancy, than the biological age when discussing the retirement timing, which is in compliance with the finding that people with chronic diseases tend to retire earlier. Currently, the simultaneous growth in longevity and mounting budget deficit in the Czech Republic have increased interest in raising the age of eligibility for retirement benefits. However, healthy life expectancy varies widely within and across gender and race groups, which deserves closer research attention (Cutler et al., 2011;Jagger et al., 2011).
Another useful approach would be the lifecourse approach reflecting different life experiences and early inequalities in education, possibilities to get a job, be promoted, make a carrier, be eligible for a good level of health care, etc. That would allow us to describe and understand the development of inequalities during people's lives and might explain the differences in the senior age. Arguing the background of aging populations and policy measures to stimulate the employment of older people, Visser et al. (2016) suggest that social inequalities in old age could grow. They are considered less-educated, and low-skilled older workers may not be able to prolong their employment career (Visser et al., 2016).
The analysis presented in this paper uses data from the SHARE database collected in 2014-2015, and the sample is representative of the Czech population. This research might serve as a knowledge base for further studies under different circumstances caused by the global pandemic COVID-19. The lockdown all over Europe and also in the Czech Republic boosted a global economic crisis and triggered unemployment, especially among the most vulnerable groups of the labour market, where definitely the seniors over 50 years old belong. Not only the age itself is currently a factor which disqualifies the older workers from the labour market, but it is also the pressure on using the digital technologies at work and also on a daily basis which might cause complications for especially older, less educated and/or less qualified employees, as well as selfemployed, precarious workers, etc. in keeping and/or finding a new job. The labour market inequalities will be stressed by the so-called digital divide (Srinuan et al., 2010). So, the importance of research focus on senior employees, and the retirement timing in the current situation affected by COVID-19 crisis is even higher.
Regarding future research of this topic, the comparison among European countries would shed light on different experiences and expectations of seniors across Europe if we realise that a growing number of EU citizens are following the work opportunities to different countries.
Another research option might also be the data analysis of a particular industry, which would improve the understanding of working seniors' specifics requirements. Unfortunately, the industry indication is often missing in SHARE data. A contributing approach would also be analysing panel data in different waves. However, not all of the variables are included in all the waves or the options offered for the answers. The language variations of some questions bring slightly different interpretations that need to be also taken into account. This particularly applies to the Look for early retirement variable, where in English, this refers to premature retirement while the Czech translation suggests a regular but early retirement. This problem may then influence the way two or more countries are compared.

CONCLUSION AND RECOMMENDATION
This paper identifies the factors influencing the decision to retire based on the SHARE database. Like many European countries, the Czech Republic faces the challenge of rapid demographic aging. Active aging policies to extend older people's working lives are the dominant response throughout Europe to keep public pensions affordable. However, the results of this study proved that retirement timing is influenced by many different factors that need to be taken into account.
A recommendation may be formulated on the government and organizational levels. The policymakers on the national level should reflect differences between gender, especially regarding different life courses for men and women. Working past age 65 should be an option rather than being enforced by raising the state pension age. There should be an option to retire at 65 or earlier on a full pension for those in physically demanding or stressful work. The issue of precarious employment needs to be addressed at the government level. The raised retirement age in private sector jobs should match the raised state pension age. Last but not least, a recommendation is regarding measures to improve the health and well-being of workers of all ages, which should be introduced.
On the organizational level, the stress should be on the implementation of age management. Among the major recommendations is an emphasis on the phased retirement, gradual retirement, and offer of bridge jobs, which means positions that can be taken by retired seniors being, for example, less physically demanding. Crucial factor increasing the wellbeing of all workers is a working environment created by companies for their employees. The aim should be ensuring the correct ergonomics in the workplace influencing the health of employees. Also, companies should care about the health of their employees, supporting their active life. Individuals can only influence some of the factors by their behaviour, but they should realize how important for them is care for their own health, lifelong learning, active lifestyle, and work-life balance.
Extended working life policies have been widely promoted in Europe and around the globe. The influence of the labour market on the hiring of older workers and their availability to work at retirement age should be examined due to the current economic downturn and digitalization.