Appendix 1: Russian old-age state pensions
The Russian pension system retained much of the structure of the Soviet pension system through the 1990s and early 2000s. Pension eligibility is based on age thresholds that were first introduced under Khrushchev’s leadership in 1956. As in the Soviet Union, old-age state pensions require only five years of work experience. The pensions had two main components including a fixed amount that is supplemented by a labor pension amount. The labor amount can vary with individual contributions based on wages, but the rate of return is prescribed by law. Base amounts were set at 55 percent of income in the preceding 2 years, plus 1 percent for every year of service beyond 25 years. Many supplements were available. For example, an additional 25 percent of the minimum pension amount was available for those with health problems. An additional two-thirds of the minimum pension amount was paid for each dependent in the household and to war veterans.
Working past retirement age was common in the Soviet Union. Pensions were reformed in 1970 to encourage this. By the early 1980s one-third of eligible pensioners continued to work. Almost all workers who delayed retirement received full pensions in addition to their income (Jones and Moskoff 1987). Except for about 10 percent of workers in white-collar occupations, everyone was able to receive pension and labor income amounts together. Even the small fraction of the civil service who are subject to age limitations in the workplace, such as public university rectors and ministers, have the option of postponing retirement for up to five years. Because pension amounts were based on past records of earnings and not indexed to inflation, replacement values remained low through the 1990s. Pension replacement rates fell from 75 percent prior to 1990 to below 30 percent by 2005 (Rashid et al. 2002; Rosstat 2015). In addition, many state-provided essentials including heat and transportation subsidies were removed. The ability of pension income to meet the needs of Russians was generally lower in the post-Soviet era than it had been just prior to the collapse.
Appendix 2: Supplementary tables and figures
Table 7 The causal impact of attaining state pension age on other subjective wellbeing outcomes. RD estimates, Russia 2006–2011
Table 8 Test for discontinuity in density of age at pension threshold
Appendix 3: Results by education and health
Estimates for sub-samples of men and women by reported health and education offer insight about the heterogeneity of effects by human capital. We split the samples according to respondent reports in 2006.
Results are estimated separately for individuals who did and did not report having a serious health problem in the 30 days prior to survey response. Table 9 presents fuzzy RD results for our preferred specification. Means of outcome variables are reported at the bottom of each panel. These means are reported separately for the periods before and after pension age, each of which span 5 years. The impact of pension receipt on the labor supply decisions of women with health problems in Panel A and without health problems in Panel B is very similar. Monthly hours worked are reduced by 26 and 21, respectively. The measured effect of pension receipt on whether or not a woman is working outside the home is negative for both groups, although it is significant only for women without health concerns.
Table 9 Main results, separately by health status in 2006 The impacts for male retirement behavior in Panels C and D are more distinct. Men reporting no health problems exhibit retirement behavior at pension age. Pension receipt decreases the likelihood of working outside the home and amounts to a reduction of 62 hours worked per month. In contrast, results are statistically insignificant among those reporting a health concern. This is true despite the fact that the impact of pension receipt on the incomes of these men with health concerns is more than twice as large as the effect on those without any health concern. Means of the outcome variables show that the men with and without health concerns work an average of 86 and 107 hours per month, respectively.
Results are also estimated separately for individuals who did and did not complete higher education. We code individuals who reported obtaining a diploma from an institute; university; post-graduate residency or individuals holding a masters degree; diploma of candidate of science or doctor of science, as having completed higher education and others as having completed less than higher education. Results in Table 10 show that only those having completed higher education exhibit significant decreases in monthly hours worked. Increases in total monthly income and life satisfaction are slightly more significant for less-educated women. These results suggest that highly educated workers are most likely to have the financial resources to retire at pension age. Those with less education must continue working. Sub-sample means show that average monthly income for less educated women is about 5,000 roubles less than for women with high education.
Table 10 Main results, separately by Education Level in 2006 For men, impacts on the probability of working outside the home are negative for both those who have and those who have not completed higher education. The impact of pension receipt on home production for cash is large and significant among those households where men have less education. Sub-sample means suggest that the larger impact for these men arises because home production output for cash was much larger in these households prior to pension age. We cannot reject a null hypothesis of no impact on life satisfaction for either sub-sample of males.
Appendix 4: Social norms
Differences in social norms arising from the historical context in Russia provide one explanation for the observed changes in home production. Age-and sex-specific norms about workers and recipients of old-age pension may help to explain wellbeing impacts if they strongly influence the perceived returns to engaging in different activities. Pension income may be expected to bring about changes in the self-perceptions and time use of older workers if pension receipt is accompanied by a retirement identity that may reflect a shift in the social norms to which an individual feels they must conform. This phenomenon has been described across social science disciplines (Clark 2003; Ekerdt 1986; Szinovacz and DeViney 1999).
The economics literature contains mixed evidence about the importance of non-monetary factors in governing behavioral responses to social transfers. Keane and Moffitt (1998), Moffitt (1983), and Ranney and Kushman (1987) suggest that welfare or food stamps might induce stigma among recipients. In the labeling effects literature, Kooreman (2000), Abeler and Marklein (2008), and Hener (2013) show that certain income sources are more likely to be spent on particular items for the household, perhaps as the result of an implicit social contract. However, Case and Deaton (1998) find no differential use of South African pensions on household expenditures relative to earned income. Impacts attributed to labelling or other non-monetary aspects of social transfers may in fact be the result of failure to control for individual-specific fixed effects or the use of IV estimates in the cross-section where exclusion restrictions are more likely violated. There is extensive evidence of social norms from the environmental economics and experimental economics literatures, including Allcott (2011), Campa and Serafinelli (2016), Cappelen et al. (2013), Dal Bó and Tervió (2013), DellaVigna et al. (2012), Fellner et al. (2013), Ferraro and Price (2013), Fershtman et al. (2012), Fischbacher and Föllmi-Heusi (2013), Gächter et al. (2013), Gneezy et al. (2016), Krupka and Weber (2009), Krupka and Weber (2013), Pruckner and Sausgruber (2013), and Viscusi et al. (2011). Social interaction effects, themselves possible manifestations of social norms have also been shown to affect participation in various programs including disability insurance in Norway (Rege et al. 2012).
Economists and psychologists have found that preferences formed during recessionary periods experienced between the ages of 18 and 25 remain remarkably stable later in life (Giuliano and Spilimbergo 2014). This suggests that social norms may be set during the sensitive ages of youth and persist through retirement age. Individuals in our sample would have formed these norms during the Soviet period.
Alesina et al. (2013) show that gender norms formed by ancestors significantly impact contemporary female labor force participation. Cigno et al. (2017) also provide complementary theoretical results which supports the transmission of preferences from parents to children. Gender norms formed during the Soviet period may also persist across generations. Lasting impressions of mothers and grandmothers may have suggested a woman’s role was to retire to the home at pension age and engage in home production (Höjestrand 2009). The observed male role, instead, was centered around the continued importance of work beyond pension age. Men generally did not retire and historically participated little in child-rearing (Utrata 2011). Thus, for older Russian men, the typical assumption that leisure is a normal good may not hold.
The moderating effects of social norms on reactions to receiving old-age pensions are likely to be particularly strong for individuals whose impressionable years were spent in the Soviet Union. Social norms can be examined using The World Values Survey (WVS 2015), a nationally-representative questionnaire designed to assess beliefs and values. Cross-sections are available for Russia for the years 1990, 1995, 2006, and 2011 and include subjective responses to questions regarding about acceptable ages for workers. Details of the questions posed are given in Appendix 5.
WVS data show that social norms about work have changed substantially across generations during the period 1990–2011. Older generations are more likely to hold norms that favor the continued employment of males beyond pension age. Meanwhile, attitudes towards work among older workers remain similar to those promoted by the state during Soviet times. Table 11 presents ordered probit estimates using existing waves of the WVS. Compared to the 1990 survey, in the 2011 survey both sexes were more likely to report agreement with the statement “When jobs are scarce, men should have more right to a job than women”. The idea that women should cede jobs to men in difficult economic times grew more acceptable over time. WVS responses indicating agreement with the statement “When jobs are scarce older people should be forced to retire” also change across generations. In 1995, the last year for which the question was posed, both sexes were significantly more likely to agree with this statement than in 1990. Agreement with these statements differs slightly between the sexes but the trend in beliefs was the same for women and men, ceteris paribus.
Table 11 Social norms about work and leisure in Russia, 1990-2011 To further demonstrate differences in social norms across cohorts, Table 12 presents the impact of reaching pension age on the subjective importance of work. Negative estimates for women suggest that they were relatively happy to retire and contribute to the family upon reaching pension age. This finding offers one explanation for increases in life satisfaction that are large relative to the mean response among women. This effect is most evident among less-educated women.
Table 12 The effect of pension age attainment on the perceived importance of work, separately by gender. Russia 1990-2011 There is no evidence that the importance of work for men decreases at pension age. The lack of change in views of men at retirement age about the importance of work to life contrasts with the observed broad changes in general views of work in Russian society. The sustained importance of work for men beyond retirement age may be partly due to persistent social norms venerating work among there older cohorts. This finding offers one explanation for increases in life satisfaction that are very small relative to the mean response among men. Additional results in Table 13 confirm that these social norms in Russia differ substantially from those held in the US.
Table 13 Comparison of views about older people in Russia and the US, World Values Surveys 1995 and 2011 Appendix 5: World Values Survey (WVS) Questionnaire
For interviews carried out in wave 3 (Russia 1995 and USA 1997), respondents were asked the following questions related to age and retirement:
-
“Do you agree or disagree with the following statement? When jobs are scarce, older people should be forced to retire from work early.”
Respondents could choose one of the following responses: Agree (1), Neither (2), Disagree (3).
-
“Please tell me how acceptable or unacceptable you think most people in [country] would find it if a suitably qualified 30 year old was appointed as their boss?”
Respondents could choose responses on a scale from Completely unacceptable (1), to Completely acceptable (10).
-
“Please tell me how acceptable or unacceptable you think most people in [country] would find it if a suitably qualified 70 year old was appointed as their boss?”
Respondents could choose responses on a scale from Completely unacceptable (1), to Completely acceptable (10).
-
“Now could you tell me whether you agree, agree strongly, disagree or disagree strongly with each of the following statements: Companies that employ young people perform better than those that employ people of different ages. ”
Respondents could choose between the following responses: Strongly agree (1), Agree(2), Disagree (3), Strongly disagree (4).
Appendix 6: Russian Longitudinal Monitoring Survey (RLMS) Questionnaire
The subjective wellbeing questions included in the RLMS 2006–2011 are:
-
(i.) “To what extent are you satisfied with your life at the present time?”
Respondents could choose one of the following responses [recoded in this paper to be increasing in satisfaction]: Fully satisfied (5), Rather satisfied (4), Both yes and no (3), Less than satisfied (2), or Not at all satisfied (1).
-
(ii.) “And now, imagine please a nine-step ladder, where on the bottom, the first step, stand the poorest people, and on the highest step, the ninth, stand the richest people. On what step are you?”
Respondents chose a number in the range 1-9, inclusive.
-
(iii.) “And now, imagine please a nine-step ladder, where on the bottom, the first step, stand the powerless people, and on the highest step, the ninth, stand the most powerful people. On what step are you?”
Respondents chose a number in the range 1-9, inclusive.
-
(iv.)“And now, another nine-step ladder where on the lowest step stand people who are absolutely not respected, and on the highest step stand those who are very respected. On which of the nine steps are you personally standing today?”
Respondents chose a number in the range 1-9, inclusive.
-
(v.) “Do you think that 12 months from now your family will live better than today, or worse?”
Respondents chose from responses: Will live much better (1), Will live somewhat better (2), Nothing will change (3), Will live somewhat worse (4), Will live much worse (5).
-
(vi.)“To what extent are you concerned about your family’s ability to procure basic necessities in the next twelve months?”
Respondents chose from responses: Very concerned (1), A little concerned (2), Not very concerned (3), Not concerned at all (4).
-
(vii.) “Tell me, please, how would you evaluate your health?”
Respondents chose from responses: Very good (1), Good (2), Average (3), Bad (4) or Very bad (5).