Note: I have made this draft paper available on the web (at www.galbithink.org) because:
1) it addresses an important issue (psychological well-being or "happiness") at a historic time and place;
2) it includes some data that is probably difficult for most scholars to find;
3) almost no one has read it, at least through Feb. 17, 2005.
This should be considered unreviewed, unpolished, incomplete work. But I hope nonetheless that it might be of some value to someone.

Psychological Self-Assessment during Radical Social Transformation: Evidence from Russia 1991-1993

Douglas Galbi

Common Security Forum

Draft: 30 November 1993

Over the last several years countries in Eastern Europe and the former Soviet Union have been trying, with varying degrees of determination, to bring about radical systemic change. The most obvious development to an outside observer has been the flourishing of personal and political freedom. Citizens have received new freedom to travel abroad, start businesses, hire labor, own property, and buy and sell land. Freedom of speech and freedom of the press, harshly repressed under communist regimes, have developed to a considerable extent. Citizens have received the right to organize political parties and vote in elections with more than one candidate. The results of these changes appear directly before one's eyes: newspapers screaming headlines about various candidates, kiosks hawking imported goods, local and foreign entrepreneurs meeting to set up new ventures.

Another significant but less visible change has been a loss of economic security for individuals. By economic security I mean one's confidence that one will have the means to follow an acceptable life plan. The loss of economic security takes several different forms. The systemic changes of the last several years have demonstrated the possibility of radical change, and have thus increased uncertainties about the future form and evolution of social institutions. Individuals are less sure about the form of the institutions to which they will have to adapt themselves. At the individual level, the guarantee of the right to a job, a cornerstone of the communist system, has been severely damaged. Individuals' places within the distribution of income are being re-arranged while the over-all spread of the distribution is increasing. Dramatic increases in inflation have increased uncertainty about exchange values and hurt individuals' ability to use savings as a means for dealing with economic risks. The real value of pensions has fluctuated significantly, increasing uncertainty about old-age income.

The effect on individuals of a loss of economic security depends in a significant sense on psychological factors. If individuals are placed in a more risky environment, one might expect to see them taking actions to limit risk factors, and to pool and redistribute remaining risks. But an increase in economic insecurity may also involve an increase in the kind of uncertainty that leaves individuals at loss as to what to do. This kind of uncertainty is associated with dread, disorientation, and hopelessness, and in the extreme passivity and fatalism. Moreover, this kind of uncertainty mitigates the effective value of personal and political freedoms: freedom has less value to individuals who are psychologically and emotionally crippled by insecurity.

This paper will analyze survey data on individuals' psychological health in Russia between 1991 and 1993. Most of the data analyzed here is from Moscow, although evidence is also presented from country-wide surveys. While the surveys from different years and sources do not use identical questions, the basic form is to ask individuals to assess their psychological state. The results indicate that, at least in Moscow, the threat of unemployment is a less significant source of stress than concerns about personal physical safety and factors associated with age and sex. Moreover, there is no evidence of growth in psychological distress since the summer of 1991. The evidence does suggest, however, that relative to other factors income differences are becoming an increasingly important determinant of mental stress.

The section will examine the correlates of psychological stress in a representative sample of Moscovites between June and August of 1991. The sample consists of 2318 observations, and was carried out by COTEKO, a joint Russian-Norwegian social research venture. The survey included questions both about objective characteristics of the respondents and their living conditions, and questions about feelings, opinions, and preferences on a wide variety of subjects.

I constructed an indicator of psychological stress from the responses to the following question:

"How often did you in the last 6 months:

a) suffer from strong heart palpitations although before this time you didn't work in a difficult or stressful job?

b) be nervous,irritable, agitated, suffer from fear?

c) feel depressed, with a loss of desire to do anything?"

In their answers to each of parts a), b), and c) above, respondents were given the choice of "often", "sometimes", and "never". Table 1 summarizes the responses given.

As my criterion for high stress in the analysis below (STRESS), I use an indicator of whether the respondent reported "often" for the existence of one or more of the stress types. The share of respondents who meet this criteria is 17%.

The question on stress seems to be capturing something more serious than general pessimism or frustration with the difficulty of life. When asked whether they were confident in tomorrow and looked on the future with optimism, 58% answered "no". When asked how they would estimate on the whole the income of their families, 68% of respondents chose "insufficient", as opposed to alternative choices of "sufficient" or "more than sufficient". Thus the share of persons who experience "often" a type of stress in Table 1 is much smaller that the share of persons who report frustration in more general terms.

A follow-up question to the stress question suggests that the kind of stress reported in Table 1 has significant behavioral effects. About 82% of the persons surveyed reported "often" or "sometimes" to one or more of the parts in Table 1. This group was then asked whether their psychological problems (nedugi nedomoganiya) limit their ability to work or associate with other people. A full 10% reported that these problems "very much" affect their ability to work or associate with people. Another 58% reported an "insignificant" effect and 33% reported "no effect".

I will examine the correlation between psychological stress and several factors. The first set of factors is demographic: age and sex. The variable AGE represents the respondents age in years, and FEM is an indicator variable with a 1 indicating a female respondent. The average age of the sampled persons is 41 years, and 56% of the sample is female.

A second factor is personal physical safety. The survey asked respondents whether recently they had been subject to an attack or the threat of an attack. Respondents were given a choice between "yes, very", "to some extent", and "no" for three different locations: on the street, in public places, and at home. The percentage reporting being "very" threatened was 19%, 13%, and 8% for the street, public places, and home, respectively. I use as an aggregate measure of personal insecurity CRIME, an indicator variable that is 1 if a respondent felt very threatened in any of the given domains. A total of 21% of interviewees felt threatened by this definition.

A third set of factors is more narrowly economic. I consider the effect of family income (INC) and whether the respondent feared losing his or her job in the immediate future (FJL). Mean family income is R910, while the percent of workers who feared losing their job is 32%.

Table 2 presents the results of a Probit analysis of the determinants of psychological stress. The coefficients indicate the effect of the independent variables on an index of the probability that a person experiences stress. One immediate observation is that personal safety (CRIME) has the greatest effect on stress. A person who reported feeling "very" threatened has a 10.5 percentage point larger probability of reporting high psychological stress. Age and sex are also significant: an increase in age of 20 years adds 7.1 percentage points to the probability of high stress, while being female adds 4.9 percentage points. Income and fear of unemployment are less significant than personal safety, age, and sex: a two-fold increase in income increases the probability of high stress by 1.8 percentage points, while fear of losing one's job adds 2.8 percentage points.

Table 2 includes two additional variables concerning personal relationships: PER-CLASH and ISOLATED. PER-CLASH is an indicator of social isolation. Interviewees were asked if they had close relationships with persons outside their family. The share of interviewees answering that they had no such person was 9.4%. The variable PER-CLASH indicates whether the interviewee reported often having misunderstanding or conflicts with one or more parties. About 14% of interviewees had such misunderstandings or conflicts. Table 3 documents the domain of the problems. The parental relationship was the most prone to misunderstandings and conflicts. In Russia parents continue to live with their children after their children marry, often in cramped accommodations. This situation may exacerbate the stressfulness of generational differences in attitudes.

Both social isolation (ISOLATED) and misunderstandings and conflicts in personal relationships (PER-CLASH) contribute to stress. Nonetheless, it is worth noting that these variables were defined using a relatively high threshold, yet their significance is only on the same order as the effect of being female. Moreover, adding these two variables has little effect on the coefficient of the other variables. Thus interpersonal conflicts do not seem to be the source of the effects being registered by the other variables.

The survey indicates that a significant share of respondents perceive that their educational level is insufficient. While 86% reported that their education level fully sufficient for carrying out their official responsibilities, only 61% judged that their education was sufficient for active participation in public affairs. Other results from this line of questioning are even more surprising. Only 43% of interviewees judged their level of education sufficient for participation and deciding economic and social issues problems in their primary workplace, and even fewer -- only 20% -- thought they had sufficient education for developing and applying decisions in the sphere of local social politics. While these results are someone suspect, and may merely reflect deference in the presence of the interviewer, it is also worth noting that 39% of the interviewees had taken some form of educational course in 1989-90, and 40% expressed a desire to raise their level of education.

However, the data do not indicate that the level of education has a significant correlation with high stress. An indicator for persons with some higher education (54% of the sample) was not statistically significant in the Probit equation reported in Table 2. Similarly, an indicator for persons with less than a middle education (less than 10 years -- 9% of interviewees) was not significantly correlated with high stress.

In summary, the evidence presented in this section indicates that personal physical security, age and sex, income, fear of job loss, and indicators of personal conflicts and social isolation were linked with high stress in Moscow in the summer of 1991. Being female was an important correlate of high stress, and the apparent impact of this characteristic was of the same order as other indicators of personal conflict and social isolation. On the other hand, the economic factors of income and fear of job loss were relatively unimportant.

II. Comparative Evidence from Russian-Wide Surveys in 1993

The results in the above section have two key limitations. First, they are based only on the population in Moscow, which may differ significantly from the population in the rest of Russia. Second, the above survey did not provide any information on change in stress over time. Given the radical social transformation occurring in Russian, such information would be particularly interesting.

Since March of 1993 the Russian-Wide Center for the Study of Public Opinion (VCIOM) has been conducting surveys which include information about interviewees' psychological state. The survey, consisted of about 4000 new interviewees each month, is representative of five macro-regions of Russia, large, medium-sized and small cities, villages, and basic demographic groups. The question relating to psychological state is: "What would you say about your mood in recent days?". Interviewees chose their response from among the following: "excellent mood", "normal, regular condition", "I'm experiencing tension, irritation", "I'm experience fear, depression", "I find it difficult to say". The results from the May and August surveys are listed in Table 4.

Table 4 includes a breakdown by the place of residence of the respondent. The category "large cities" includes Moscow, and hence it is the category most comparable with the data in the previous section. There appears to be no strong, consistent pattern over time of differences across types of locations. I focus on the share of respondents who reported experiencing fear or depression. In Table 4 the share reporting this condition was smallest in large cities in both months. But in July the shares for large cities, medium/small cities, and rural was 9.8%, 10.7%, and 8.5% respectively. Thus the share experiencing fear or depression was larger in large cities than in the country-side. There is some indication, however, that the share of persons experiencing fear or depression is generally smaller in large cities than in medium/small cities. This suggests that the results in the previous section, drawn from Moscow alone, might tend to under-estimate the share of persons experiencing psychological trauma in Russia as a whole in the summer of 1991.

Table 5 gives the share of respondents experiencing fear or depression among all those surveyed in each month of 1993. The share shows no evidence of a trend over time, and appears to fluctuate about a share of 10%. A comparison over a longer period of time would be more interesting, and with some caution one can compare these data to the data in the 1991 survey (Table 1). In addition to the difference in geographical coverage, the data in Table 5 relate to the respondents' experience over the past few days, while the data in Table 1 concern the respondents' experience over the past six months. Perhaps the best way to handle the comparision is to focus on those with acute problems. For each problem surveyed in Table 1, the share experiencing it often is 10-12%, and the share having often had one of these problems is 17%. Thus the evidence suggests that the incidence of acute psychological distress has not increased in Russia since the summer of 1991.

It is certainly not the case that there was a lack of potential sources of psychological trauma between the two surveys. The first survey, which took place between June and August of 1991, concluded in the month of the attempted coup. There was mass demonstrations in Moscow as tanks rolled into the city. Boris Yeltsin became a hero with his daring defence of the White House, and the Soviet Union dissolved. A new radical program of economic reform was pushed forward. In common conversation this new economic program was referred to as "shock therapy" -- the name for a form of psychiatric treatment that involves inducing a coma to lessen the pain of ensuing therapeutic electrical joints to the patient's body. The last of the second set of surveys took place in August of 1993, during a period of increasingly bitter conflict between the Russian Parliament and President Yeltsin. The depth and seriousness of the problem was underscored by the events of the next two months. Yeltsin abolished the Parliament, and when some members and their supporters refused to leave the building, tanks and troops assaulted it, turning the White House into burning, half-black symbol of national disorder.

One might have expected these dramatic events to have registered a larger psychological toll. It may be the case that 70 years of Communist Party rule taught the Russian public to maintain psychological distance from public life, a strategy for maintaining personal security in the face of political powerlessness. But such a theory does not explain how Russians have coped with the dramatic economic changes that have directly affected daily life. Perhaps in the context of the dislocations remembered in the ever-vivid events of the Great Patriotic War, to the Russian recent events look less dramatic than they do to an outsider. While that may be the case, the next section will show that there are other, more subtle psychological effects of the social transformation of recent years.

One would like to know not just whether the over-all level of acute psychological stress has increased, but also whether the correlates of stress have changed. The tabular summaries of the 1993 VCIOM survey are not sufficient for detailed analysis of this issue, and I have not yet obtained access to the raw data. There exists, however, a COTECO survey in November of 1992 consisting of a sample of 899 Moscow residents drawn from the pool of persons sampled in the summer of 1991. This section will compare the correlates of psychological stress in that survey to those found in the analysis of the 1991 survey (Section I).

The November 1992 COTECO survey approached the issue of psychological stress in the following way. Respondents were asked: "How would you evaluate your state of health?". Respondents then had to rate both their "physical health" and "nervous/mental health" on a five-point scale ranging from "very bad" to "excellent". Table 5 gives the results from "nervous/mental health"; ratings for physical health were very similar. The close link between judgements of physical and mental health may reflect a intrinsic connection between physical and mental health as well as the psychological component of a self-assessment of one's general physical health. In the analysis which follows I will take as my indicator of high stress (STRESS) those respondents who reported "very bad" or "bad" for their nervous/mental health.

The correlates of stress available in the November 1992 survey are a subset of those from the summer 1991 survey. Age and sex variables are included as before. A crime vulnerability variable is not available in the November 1992 data. However, omitting it from the analysis does not affect the values of the other coefficients in Table 2, hence its omission may not affect the comparability of the other coefficients across time. The fear-of-job-loss (FJL) variable is almost the same as that in the previous survey. However, in this survey 23% reported that there was a threat that they would lose their job in the near future, as compared to 32% in the previous survey. The income variable (INC), as before, is a measure of family income from all sources. The distribution of income among the sample in November of 1992 is much more inequal that it was in the summer of 1991. The upper tail of the income distribution was much extended in 1992, driving the ratio of the standard deviation to the mean up to 2.6 in November 1992 from being about one in summer 1991. Moreover, the growth of inequality was not just a phenomenon of the upper tail: while the ratio of the interquartile range to the median in the summer of 1991 was .8, in November of 1992 it was 1.4.

Table 6 presents a Probit analysis of the correlates of stress similar to that provided for the summer 1991 data in Table 2. Given the differences in the definition of the underlying dependent variables, the results in Table 2 and Table 6 cannot be directly compared. Note, however, that the age effects in the two tables are about the same, while the income effect found in November 1992 is about three times as big as that in the summer of 1991. This result suggests that income is becoming an increasingly important determinant of psychological stress.

I can think of two tentative explanations for why income may be becoming a more important determinant of psychological stress. First, the growth of income inequality may be raising the psychological profile of income differences. Income differences are becoming more obvious in society, they are conflicting with and challenging older forms of social status, and those with low incomes are becoming more acutely aware of their relatively disadvantaged position. Second, the government's commitment to support even a minimal standard of living may be becoming increasingly suspect. As food prices are liberalized, living accommodations privatized and marketized, and social services neglected, economic security, and hence psychological stress, may be increasingly associated with personal money income.

Several important points emerge from the above evidence. First, about 10% of Russians report experiencing fear, depression, or other indications of a troubled mental state. The incidence of these kind of problems is correlated with crime and demographic factors such as age and sex, while the fear of unemployment does not register as a major source of stress. There is no evidence that the incidence of acute mental stress has increased during the past two years of radical social change. There is some evidence, however, that income is becoming more important, relative to other more diffuse factors, in determining stress. This suggests that increased income equality may begin to have an important effect on the psychological stress that individuals' experience.

Much of the analysis of public mood is concerned with the possibility of social unrest and public disturbances. One hears of "social tensions rising" and "the dangers of a social explosion". Countering these ominous warnings are assertions about the ability of Russians to endure suffering and about the extraordinary patience of the Russian people. What is missing in these accounts is attention to the effects of psychological stress on the welfare and functioning of individuals. This paper has made a start on the latter question by analyzing available evidence on psychological stress during a particularly dramatic period of change in Russia.