Presentation by Jamele Rigolini, Senior Economist in the World Bank’s Office of the Chief Economist for Latin America and the Caribbean, on 19 april , at SAE.
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
IPC IG Seminar- SAE : "Middle class values"
1. Is there such thing as middle class values?
Class differences, values and political
orientations in Latin America
Luis-Felipe Lopez-Calva
Jamele Rigolini
Florencia Torche
The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of
the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World
Bank or the governments they represent.
2. Context
• Middle class values have long been perceived as
drivers of social cohesion and growth
• There is, however, very little solid evidence
• In this paper we investigate the relation between
class, values, and political orientations
– We use Ecosocial surveys
– We measure class using households’ permanent income
– We do check robustness with occupation
3. Main Findings
• Permanent income is robustly associated with
values
– It remains so when correcting for individual characteristics
• Overall, we do not find strong evidence of any
“middle class particularism”
– values appear to gradually shift with income
– If any, the only peculiarity of middle class values is
moderation
• Changes in values across countries are much
larger than the ones dictated by income
– individual values vary primarily within bounds dictated by
each society
4. Literature: Postulates
• The middle class is the source of economic values that
emphasize savings and accumulation of human capital
– Weber (1905); Murphy et al. (1989); Matsuyama (1992 and 2002);
Birdsall (2000); Doepke and Zilibotti (2007)
• The middle class has also been associated with political
stability and social cohesion
– Aristoteles [Politics]: It is possible for those states to be well governed
that are of the kind in which the middle-class is numerous […] for by
throwing in its weights it sways the balance and prevents the opposite
extremes from coming into existence”
– Lipset (1959); Benhabib and Rustichini (1996); Barro (1999); Benhabib
and Przeworski (2006)
5. Literature: Empirics
• The empirical literature on middle class values remains however
scant
– Lack of accurate income data and sampling issues in values surveys
– Lack of consensus about middle class definition
– Reverse causality and omitted variable biases
• Self reported middle class status:
– Amoranto, Chun and Deolalikar (2010 – WVS); OECD (2001 –
Latinobarometro)
– Status is scattered across income and relates poorly to it
• Status based on income:
– Pew Research Center (2011 – mostly summary statistics); Cárdenas,
Kharas and Henao (2011 – large income brackets for Peru)
• Our analysis also relates to literature on status and redistributive
preferences
– Tóth and Keller (2011), Reed-Arthurs and Sherin (2010)
6. Values Survey
• We use the 2007 Ecosocial surveys for six Latin
American countries
– Representative or urban areas
– We only include people aged 25+
– We exclude Argentina because of no matching measure of
permanent income
– Total of 12,297 observations
7. Income Measure
• We use information of households’ assets to
construct a measure of households’ permanent
income
– We match assets & HH characteristics in Ecosocial with the
ones of an “external” survey
– An additional challenge consists in identifying respondent
in Ecosocial to match household head characteristics
– We run an income regression using an external household
survey and retrieve the “Betas”
– We impute permanent income in Ecosocial using the
retrieved “Betas”
8. Class Measures
1. Log income
2. Absolute measure
– 0-4 $/day (per capita PPP); 4-10$/day; 10-20 $/day;
20+ $/day
3. Relative measure
– Based on EGR polarization criterion that minimizes
within group and maximizes across group inequality
4. We also test robustness with respect to
occupation
9. Values Indexes
• We select first a series of survey questions capturing
orientations that are plausibly related with each other
• We then extract the weights of each variable in the first
principal component and impute it to each observation
– We discard variables presenting a high uniqueness
– With PC, the idiosyncratic determinants of each individual
item (and misunderstanding of a particular survey
question) are ruled out
• Example: “trust in institutions” is based on trust in the
national government, congress, political parties, the mayor,
and the police
10. Values Indexes
• Trust in institutions
– Index based on how much individuals trust (a) the government, (b)
congress, (c) political parties, (d) the mayor and (e) the police
• Political alienation
– Index based on respondents’ agreement with the following
statements: (a) people who are in charge do not care about people like
me, (b) the authorities would not do anything if there was a serious
problem in my neighborhood, and (c) most people in power only try to
take advantage of people like me
• Perception of mobility and opportunity:
– Index based on the following indicators: (a) How likely is for average
youth to graduate high school? (b) How likely is for a poor person to
overcome poverty? (c) How likely is for a person to start their own
business? (d) How likely is that a smart youth without economic
resources to enter college? (e) How likely is that a woman attains a
good job? (f) How likely is that any worker becomes a homeowner in a
reasonable time? (g) In this country, anyone who works hard can get
ahead…
11. Values Indexes
• Support for individual rights
– Indicator based on the survey question “Which one of these
statements do you agree with the most: Individual rights should be
respected under any circumstances (coded 1), or criminals should not
have the same rights as honest people (coded 0)”
• Legitimization of political violence
– Index based on the following items: Do you think the use violence is
justified in the following cases? (a) When indigenous minorities claim
their ancestral land, (b) When revolutionary social change is
attempted, (c) When the environment is being protected, (d) When
the poor claim better living conditions, (e) When people oppose a
dictatorship
• Voting
– Index based on the survey question: Do you vote in presidential
elections, recoded so that 1=always vote, 0=sometimes or never vote
12. Values Indexes
• Social tolerance
– Index based on the following indicators: How uncomfortable would
you be in the following situations? (a) If your child married someone
of a class lower than yours, (b) If your neighbor was a different race,
(c) If your child had a homosexual friend, (d) If your neighbor was an
immigrant worker, (e) If your neighbor was of a class lower than yours,
(f) If your child married an atheist
• Nationalism
– Index based on the following indicators: (a) Considering the good and
the bad, I am proud of being (nationality), (b) (Country) should defend
our national interests, even if that creates conflicts with other
countries, (c) (Country) should limit imports of foreign product to
protect the national economy, (d) TV in (country) should favor national
shows and films.
• Political ideology
– Indicator based on the survey question “Using a scale in which 1
means a left-wing position and 10 means a right-wing position, where
would you place yourself?”
13. Values Indexes
• Interpersonal trust
– Indicator based on the survey question “Speaking in general, do you
think most people can be trusted (coded 1) or you can never be too
careful (coded 0)?”
• Interpersonal alienation
– Index based on the following indicators: (a) In general, what I think
does not count very much, (b) I am always left out of things going on
around me, (c) People around me would not do much if something
happened to me, (d) Most people try to take advantage of me
14. Descriptive Statistics
Variable Mean sd min max N
Male 0.47 0.50 0 1 12297
Age 44.33 14.01 25 97 12285
Race/ethnicity: 3.01 1.34 1 4 12104
White 28.7%
Black 3.2%
Indigenous 6.4%
Mixed 61.8%
Mother’s education 2.99 2.41 1 10 10611
Father’s education 3.40 2.78 1 10 9979
R’s education 5.01 2.88 1 10 12289
15. Social Class Distribution
Absolute Relative
Poor (0-4 $ a day) 13.6% 45.8%
Vulnerable (4-10 $ a day) 43.9%
Lower-middle Class (10-20 $ a day) 27.3% 35.1%
Upper-middle class (20+ $ a day) 15.2% 19.1%
Total 100% 100%
16. Relative MC: Thresholds by Country
• Strong overlap of Relative and Absolute middle
class measures
Lower threshold Upper threshold
($ a day) ($ a day)
Brazil 9.0 17.5
Chile 11.9 22.3
Colombia 7.5 16.8
Guatemala 7.0 17.6
Mexico 7.4 15.8
Peru 6.2 12.6
19. Trust in institutions Political alienation
0.30 0.30
0.25 0.25
0.20 0.152 0.20
0.15 0.15
0.10 0.065 0.10 0.049 0.049
0.034
0.05 0.012 0.05
0.00 0.00
-0.05 -0.05 -0.021
-0.10 -0.10
-0.15 -0.15
-0.134
-0.20 -0.20
-0.25 -0.25
-0.30 -0.30
Education Vulnerable Lower-middle Upper-middle Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class (0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day) (10-20 $ a day) (20+ $ a day)
Support of individual rights under any Perception of opportunity
0.30 circumstances
0.30
0.25
0.25
0.20 0.20 0.176
0.150
0.15 0.15 0.121
0.10 0.053 0.10 0.053
0.05 0.05
0.00 0.00
-0.05 -0.05
-0.10 -0.065 -0.10
-0.15 -0.15
-0.120 -0.120
-0.20 -0.20
-0.25 -0.25
-0.30
-0.30
Education Vulnerable Lower-middle Upper-middle
Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class
(0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day)
(10-20 $ a day) (20+ $ a day)
20. Legitimization of political violence Voted
0.30 0.30
0.25 0.25 0.223
0.20 0.20 0.175
0.150
0.15 0.15 0.098
0.10 0.10
0.05 0.05
0.00 0.00
-0.05 -0.05
-0.032
-0.10 -0.072 -0.10
-0.15 -0.103 -0.15
-0.132
-0.20 -0.20
-0.25 -0.25
-0.30 -0.30
Education Vulnerable Lower-middle Upper-middle Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class (0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day) (10-20 $ a day) (20+ $ a day)
Social tolerance Nationalism
0.30 0.30
0.25 0.25
0.20 0.20
0.15 0.099 0.15
0.10 0.067 0.10
0.036
0.05 0.006 0.05
0.00 0.00
-0.05 -0.05 -0.003
-0.10 -0.050 -0.10 -0.043
-0.15 -0.15 -0.096
-0.20 -0.20
-0.25 -0.25
-0.30 -0.30
Education Vulnerable Lower-middle Upper-middle Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class (0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day) (10-20 $ a day) (20+ $ a day)
21. Left-right political ideology (1=left, 10=right) Interpersonal trust
0.30 0.30
0.25 0.25
0.20 0.20
0.141
0.15 0.15
0.081
0.10 0.059 0.10
0.05 0.010 0.05 0.009
0.00 0.00
-0.05 -0.05 -0.006
-0.028
-0.10 -0.047 -0.10
-0.15 -0.15
-0.20 -0.20
-0.25 -0.25
-0.30 -0.30
Education Vulnerable Lower-middle Upper-middle Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class (0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day) (10-20 $ a day) (20+ $ a day)
Interpersonal alienation
0.30
0.25
0.20
0.15
0.10
0.05
0.00
-0.05
-0.10 -0.054
-0.15
-0.20 -0.163 -0.170
-0.25
-0.30 -0.272
Education Vulnerable Lower-middle Upper-middle
(0-4 $ a day) class class
(10-20 $ a day) (20+ $ a day)
22. Trust in Institutions Perceptions of opportunity
0.80 0.80
0.60 0.60
0.40 0.311 0.40
0.189 0.226
0.166
0.20 0.20 0.114
0.059 0.040
0.00 0.00
-0.20 -0.061 -0.20
-0.194 -0.228
-0.40 -0.40 -0.331
-0.60 -0.60 -0.526
-0.80 -0.80
Ln(income) Brazil Chile Colombia Mexico Peru Ln(income) Brazil Chile Colombia Mexico Peru
Social tolerance Legitimization of political violence
0.80 0.80
0.60 0.60
0.40 0.288 0.40
0.259
0.20 0.108 0.20 0.078
0.015
0.00 0.00
-0.015 -0.046
-0.20 -0.20
-0.263 -0.232
-0.40 -0.40
-0.60 -0.60 -0.472
-0.526
-0.80 -0.80
Ln(income) Brazil Chile Colombia Mexico Peru Ln(income) -0.774
Brazil Chile Colombia Mexico Peru
23. Conclusions
• Log income is robustly associated with values
– Often (but not always) in line with education, though
effect is in addition to education
– Much more robustly so than occupations (not shown)
• We do not find however strong evidence of
special “middle class values”
– If any, middle class values show moderation and lay
between the ones of the poor and the rich
• Income (and occupation) explain much less
than other variables (country effects)
– Overall R squared remains very low