Numerous studies have shown a U-shaped relationship between age and the satisfaction level. This is explained by the fact that at a certain stage of life occur such significant life events as wedding, birth of children, retirement and others (Frey and Stutzer 2002). In the work of Andrew Clark (Clark 2007) is provided an analysis of how belonging to a certain generation can influence life satisfaction. Basing on the panel data of 14 waves of the British Household Panel Survey (BHPS), the author has built a fixed-effects model and a pooled regression. When controlling for cohort effects, the researcher identified a U-shaped relationship between life satisfaction and age. In thier article, Wencke Gvozdz and Alfonso Sousa-Poza (Gwozdz and Sousa-Poza 2010) tested the hypothesis about the relationship of age and life satisfaction based on the German Socio-Economic Panel (GSOEP). The researchers revealed a U-shaped dependence: people aged 16–19 and 70–74 are most satisfied with life, and in older ages the satisfaction level decreases significantly. However, when fixed effects were included, the influence of age was neutralized. Life satisfaction is stable over time, and it only declines among the population aged over 85.

In the work of Rosa Kutubaeva (Kutubaeva 2019), life satisfaction is analyzed in three countries – Sweden, Austria and Spain. This study bases on the panel data from four waves of the Survey of Health, Ageing and Retirement in Europe (SHARE 2018). Satisfaction is graded on a scale from 0 – “total dissatisfaction” to 10 – “total satisfaction”. The study was conducted separately for men and women. The author built several models – ordinary OLS and panel regressions with fixed effects (with and without control variables). As a result of the analysis, the age effect was confirmed only for Sweden. In other countries, a stable relationship between age and life satisfaction has not been identified.

Many studies hypothesize the relationship between place of residence and life satisfaction in older ages. In the work of Salman Zaidi and co-authors (Zaidi et al. 2009), the authors concluded that residents of metropolitan areas are less happy than residents of small towns.

Studying the effects of health on life satisfaction level is of particular interest to many researchers. The analysis of such relationship is highly relevant when assessing the level of life satisfaction of the older population. In their work, Anne Berg and co-authors (Berg et al. 2006) assessed the level of life satisfaction among the population over the age of 80. It turned out that having a diagnosed disease did not have a significant impact on life satisfaction level. However, self-assessment of individual health affects subjective well-being. The study by Ann Bowling and co-authors (Bowling et al. 1996) revealed that self-assessment of health, as a subjective factor of health status, is one of the main factors influencing life satisfaction level. The same relationship has been found for populations aged 65–89 with reduced self-care ability (Borg et al. 2006). In Europe, health plays an important role for older populations. In the study of Rosa Kutubaeva (Kutubaeva 2019), self-assessment of health turned out to be a significant variable which positively affects individual subjective well-being in all models. Similar results were obtained in the study by Wencke Gvozdz and Alfonso Sousa-Poza (Gwozdz and Sousa-Poza 2010).

Most studies use factors such as gender, education, having a job, income and family status as control variables while estimating the level of life satisfaction. These factors alone may not be significant. However, in a large number of studies, separate analysis of results, for example for men and women, can show different results. Thus, Silvia Meggiolaro and Fausta Ongaro (Meggiolaro and Ongaro 2015) used panel data coming from two waves of the Aspects of Daily Life survey conducted by ISTAT in their research, and estimated linear regressions controlling for different factors. It turned out that in Italy, for the population aged over 65, the high level of education is equally important for men and women, but the parameter becomes insignificant for women when controlled for health status. Also, the authors revealed that older couples living apart from their children feel happier than those living with them. Some studies (Takashi 2012) indicate that for women family status is more important than for men. At the same time, men, after getting divorced or widowed, rate their well-being lower. However, estimates of the family status impact for both genders show that it is a significant factor. Generally, people living with partners show higher levels of life satisfaction (Buber and Engelhardt 2008). Gennady Voronin and co-authors (Voronin et al. 2018) conducted an analysis of the life satisfaction among individuals older than working age living without a family or a partner on the basis of the questionnaires of the Russian Longitudinal Monitoring Survey – Higher School of Economics (RLMS-HSE). The elderly living alone were found to be less satisfied with life than those living with family, and their incomes were lower. Both single men and single women who reach retirement age strive to continue their work activities longer than those who have families. Single women are more likely to have chronic diseases and evaluate their health lower than single men.

A number of studies have shown that income also has a positive impact on the life satisfaction levels of older people (Fernández-Ballesteros et al. 2001). However, the scale of such influence varies across the studies. The study by Martin Pinquart and Silvia Sörensen (Pinquart and Sörensen 2000) identified a weak relationship between income and life satisfaction. It is assumed that, over age, people often limit their needs and adjust them to their current financial capacity, which weakens the relationship between income and subjective wellbeing.

The question of how much life satisfaction depends on the employment status has interested many researchers. The study by Yanni Hao (Hao 2008) analyzes the impact of having a job on the rate of mental state decline among people of older age. The study was conducted in the United States basing on the Health and Retirement Study (HRS), a longitudinal survey of the population of groups born in 1931–1941. The author considered individuals who took part in all rounds of the survey from 1996 to 2002 (the survey was conducted every 2 years). A growth curve model was used to conduct regression analysis. According to the results of the study, employment has a positive effect on the mental state of the individual. The author compared three types of employment: full employment, partial employment, and volunteerism. At the first round of the survey, there was no significant difference in the life satisfaction levels between individuals with different types of employment. However, over time, those who worked full-time showed slower mental health decline. Participation in volunteering also has a positive effect on well-being, although to a lesser extent than work.

In the study by Andrew Clark and co-authors (Clark et al. 1996), the authors identified a U-shaped relationship between job satisfaction and age. For older persons, together with regular income in the form of wages, the availability of work also has significant non-monetary benefits such as public recognition and social connections. Also, older men often hold senior positions. However, benefits may vary by the type of employment. Those who work full-time are more likely to receive bonuses, as well as lucrative benefits in the form of health and pension insurance, compared to those who work part-time (Bosse 1990).

However, some studies have found an inverse relationship between life satisfaction and work (Wallace 2008). Older non-working individuals feel higher levels of life satisfaction compared to those who continue to work. Perhaps this is due to the type of activity of the individuals under review.

One of the local studies of the life satisfaction level of older persons (Kolosnitsyna et al. 2017) analyzes the influence of different groups of factors on the subjective well-being of individuals of retirement age. The authors hypothesized that life satisfaction of Russians in older ages depends on the following variables: gender, age, place of residence (type of settlement), health, education, income, employment, social status and family characteristics.

The study was conducted on the basis of panel data of the Russian Longitudinal Monitoring Survey – Higher School of Economics (RLMS-HSE), rounds 18–21. Russians aged over 55 were included in the sample. Ordered logit models, as well as OLS models with fixed and random effects were used to obtain empirical results. According to the results of the study, belonging to a particular gender is not a significant variable itself, however, gender factors to some extent affect the well-being of men and women. In Russia, married men tend to be happier than married women while ageing. Having children for men is not a significant factor, and for women it has a weak and negative effect on subjective well-being. In other studies (Buber and Engelhardt 2008), older men living in Europe and having up to three children were proved to have significantly lower rates of depression compared to those with no children or with four and more children.

For older women, the most important factors are employment status, education and health self-assessment. The level of education is significant for women and has a positive impact on their well-being, while for men the factor has a weak negative effect. With age, working women also feel much happier than their male counterparts. Perhaps this is because, for women in older ages, work is not so much about income as it is about being able to communicate with others and build social relations.

Thus, the literature review identified the following groups of factors potentially determining the level of life satisfaction of the older population: (1) age, gender and place of residence; (2) health and its self-assessment by individuals; (3) characteristics of socioeconomic status: education, income, employment status, type of occupation; (4) family and family bonds defined by the number of children, the marriage status and the number of household members.

The following hypotheses are tested within this study:

1. Having a job has a significant positive impact on life satisfaction of the older population.
2. There are differences in the life satisfaction level between individuals with different occupational status.
3. The presence of a pension as a source of income has a positive impact on life satisfaction in older ages.

This study is based on examining the impact of the individual employment characteristics on the level of life satisfaction of older persons. This choice is justified by the observed gradual increase in the labour supply among the population aged over 45. The pension reform is expected to enhance this trend in the coming years, which is why employment status may be one of the most significant factors influencing the subjective assessment of the individual well-being. Besides, the authors suggest that the type of occupation may have an additional significant impact on the level of life satisfaction due to the existing differences in the type of job and labour conditions, as well as in the type of employment.

Thus, the distinguishing feature of this study is the inclusion of the employees’ qualification level.

Given the panel structure of the data, we estimate the OLS model with fixed and random effects first, along with a pooled regression. Hausman test showed that the model with fixed effects was more preferable, and the results of its estimation are presented in Table 2. Since residence and gender remain unchanged in time, the effect of their influence is attributed to fixed effects.

Estimates of the OLS model with fixed effects show that such indicators as marriage and employment status, pension and income have a significant positive impact on life satisfaction level among older persons of both sexes. Health self-assessment also significantly affects life satisfaction – life satisfaction level decreases with the decline in health. The coefficients estimates for variables age and age squared are not statistically significant, suggesting that age has no effect on life satisfaction. This may be explained by the use of fixed effects, which include the factor of belonging to a particular generation and neutralize the influence of age.

Estimating the same model separately for men and women reveals the following differences. Unlike men, for whom family status is not a statistically significant factor, for women marital status has a positive impact on life satisfaction. Also having a child in the family makes a woman happier, while the influence of this factor on life satisfaction among men is absent. Most likely, the insignificance of the variables standing for family status and having children for men might be explained by the fact that in Russia the traditional family pattern, in which women pay more attention to family and household, is predominantly preserved. Disability has a significant negative impact on life satisfaction level for men, when for women we do not observe this correlation. According to our assumption, this may also be related to patriarchalism – the official disability status can negatively affect men’s activities, reduce their business activity and lead to difficulties in maintaining the status of the head of the family. The education level turns out to be significant at a 0.1 level, but only in the male sub-sample, and this factor has a weak negative effect on individual life satisfaction.

The literature analysis showed, that quite a number of authors use probability models since in some cases they are better suited for population surveys. Taking into account the panel data structure, we have estimated binary logit models with fixed effects for each gender. The dependent variable life satisfaction was transformed into a binary form, where 0 is attributed to individuals satisfied with their life (answers 1, 2 or 3), and 1 to those not satisfied (answers 4 and 5). The estimates of this model are given in Table 3.

In the logit model with fixed effects, statistically significant variables are marital and employment status, availability of pension and income level; all of them positively affect life satisfaction level. In addition, when assessing the sample for both sexes, we find the number of household members to be significant at a 10% level – the more members of one household live together, the higher is the level of life satisfaction of the respondent. However, this result is not robust; when considering each gender individually, we find this factor to be statistically insignificant.

Differences in the results obtained for the logit model with fixed effects estimated separately for men and women are partly similar to those of the OLS model – marital status is insignificant in the male sub-sample, and having children has a positive impact on women’s life satisfaction. In addition, the employment status loses its significance for men, which we associate with the possible error growth due to the shift to a binary model.

In order to determine the health level of individuals, as well as their level of life satisfaction, we examine the indicators subjectively assessed by each of the respondents. Using health self-assessment as a proxy-variable can produce biased estimates of real health level. However, the use of a subjective indicator enables seeing direction and assessing the importance of health influence on the level of well-being of older individuals.

The transformation of a categorical dependent variable into a binary one for the logit regression could lead to a decrease in the accuracy of the obtained results. Within the current analysis, the dependence of life satisfaction level on major significant factors persists in different specifications, including those with a binary variable. However, in future studies, ordered logit model can be constructed to preserve the discreteness of the dependent variable.

The study does not take into account the frequency of changing individual qualifications. However, this indicator is of great interest since the concept of lifelong learning is now spreading actively. Continuing education is provided by the state by creating opportunities for the development of competences and changing qualifications at any age. Such changes in the labour sector can have a significant impact on the subjective level of individual well-being. Therefore, studying the influence of the frequency of changing qualifications may be the next step in analyzing older people’s life satisfaction.