This document presents a study examining the influence of one's childhood economic environment on later-life risk aversion. The study uses data on risk attitudes and birth dates from the Health and Retirement Study. Preliminary ordered probit models find that individuals born during World War II report higher risk aversion, while childhood education levels and economic periods have mixed effects. The author discusses extensions, such as analyzing how age during economic crises impacts risk preferences over the life course.

1. Childhood’s Economic Framwork
and its Effects on Preferences
Melika Liporace
HEC Lausanne
April 17, 2014
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 1 / 16

2. Outline
1 Introduction
2 Data
Risk Aversion
Birth Date
3 Methodology
Assumption
Control Variables
4 Preliminary results
Control Variables
Significance
5 Conclusions
Extensions
Final words and debate
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 2 / 16

3. Introduction
Introduction
Main question:
Does the economic framework of one’s childhood influences one’s
risk aversion during later life?
Motivation:
HRS Data naturally offers a lot of information on effects of economic
changes through history
Risk Aversion is often ruled out as an heterogeneous feature, and not
really analyzed
Childhood’s Economic Framework can be seen almost as a random
variable, reducing the risks of endogeneity often encountered in
preferences and microbehavioral analysis
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 3 / 16

4. Data Risk Aversion
Data: Risk Aversion
Design
Two questions are asked:
Suppose that you are the only income earner in the family, and you
have a good job guaranteed to give you your current (family) income
every year for life. You are given the opportunity to take a new and
equally good job, with a 50-50 chance it will double your (family)
income and a 50-50 chance that it will cut your (family) income by a
third. Would you take the new job?
Yes: Suppose the chances were 50-50 that it would double your (family)
income, and 50-50 that it would cut it in half. Would you still take the
new job?
No: Suppose the chances were 50-50 that it would double your (family)
income and 50-50 that it would cut it by 20 percent. Would you then
take the new job?
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 4 / 16

5. Data Risk Aversion
Data: Risk Aversion
Design
4 categories:
Risk Aversion Type Answer Q1 Answer Q2
I Yes Yes (2a)
II Yes No (2a)
III No Yes (2b)
IV No No (2b)
Issue: from Wave 4 and forward:
Two additionnal questions for Yes/Yes and No/No ⇒ two additional
categories ⇒ ”rescale”:
Wave 4 type 1 2 3 4 5 6
Wave 1 type 1 1 2 3 4 4
Question rephrased to avoid ”status-quo bias”
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 5 / 16

6. Data Risk Aversion
Data: Risk Aversion
Summary Stastics: relative frequencies
Relative Freq.
1 12.60
2 10.26
3 13.60
4 63.55
Total 100.00
0204060
Percent
0 1 2 3 4
Risk Aversion
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 6 / 16

7. Data Risk Aversion
Data: Risk Aversion
Relevance
Pros:
Measured by a specific design as well adapted as possible
No utility specification assumed
Gamble over lifetime income (vs. usually, small bets)
Barsky et al. (1997)1
found that ”The measured risk tolerance has
predictive power for choices over risky behaviors - the decisions to
smoke and drink, to buy insurance, to immigrate, to be self-employed,
and to hold stock”
Cons:
Self-reported choices
Hypothetical situations
Status-quo bias
Mitgated findings of Barsky et al.: ”the incremental predictive power of
risk tolerance is never very high.”
1
Preference Paramters and Behavioral Heterogeneity: an Experimental Approach In Health and Retirment Study
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 7 / 16

8. Data Birth Date
Data: Birth Date
Summary Statistics: Frequency
0200400600800
Frequency
1900 1910 1920 1930 1940 1950 1960 1970 1980
Birth Date
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 8 / 16

9. Data Birth Date
Data: Birth Date
Economic era: dummies
3 main economic periods
Great Depression (1929-1939)
World War II (1939-1945)
After War, Baby-boomers (after 1945)
Rule: Period’s dummy is 1 if: at least 10-year-old at the end of the period
Frequency % of pop.
Great Depression Dummy 2158 0.1325
World War II Dummy 5650 0.3468
Observations 16292
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 9 / 16

10. Data Birth Date
Data: Birth Date
Economic era: dummies
Possible issues with data:
Small sample: ”Few” people in the Great Depression category
Sample selection: Individuals survived because more risk averse
Note:
Base group: Babyboomers
(actually means have known neither period)
Caution with WWII dummy
(actually means have not known Great Depression)
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 10 / 16

11. Methodology Assumption
Methodology: Assumption and Specification
Main assumption: Risk Aversion is time invariant:
Preferences are shaped during childhood and do not change after that.
⇒ Age per se should have no influence on it
⇒ Age influences risk aversion only through the economic history
⇒ Same reasoning for almost every other potential control variables
Then (and only then): almost no risk of endogeneity
Model: ordered probit (logit), with:
Latent variable riskavers∗
= ecodummies β + zγ + ε
Outcome riskavers ∈ (0; 4), as the usual function of y∗
Interested in partial effects of ecodummies (partially through β)
Control variables z
→ which control variables ?
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 11 / 16

12. Methodology Control Variables
Methodology: Control Variables
Idea : find variables that:
are correlated to individual’s birth date
influence risk aversion (through other channels than economic past)
Two main candidates:
Education:
Birth date changes access and average education, which shapes risk aversion.
(Arbitrary variable choice: years of education)
Parent’s education:
risk aversion could be partly inherited through upbringing
parent’s education could be a (largely imperfect) proxy of parent’s
income, therefore the encountered effect of economic crisis
(Only available variable: level of education)
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 12 / 16

13. Preliminary results Control Variables
Preliminary results: Control Variables
Mean
Years of Education 12.60
Parent’s Education 10.33
Observations 14802
Control variables requirements :
drops less than 2000
Risk aversion and Economic
dummies distribution roughly
stay the same
⇒ Not too demanding
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 13 / 16

14. Preliminary results Significance
Preliminary results: Significance
Table 1: Basic Ordered Probit Model
Risk
Great Depression Dummy 0.0533
(0.0367)
World War II Dummy 0.0982∗∗∗
(0.0253)
Parent’s education -0.00109
(0.00392)
Years of Education -0.0164∗∗∗
(0.00320)
Standard errors in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
In economic dummies:
only WWII dummy is
significant
Surprise
parent’s education very
significant
education not
significant at all
⇒ Should we collude both
dummies in only one such
as ”Before/after WWII”?
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 14 / 16

15. Conclusions Extensions
Extensions
Further analysis:
Does the age at which one has encountered an economic crisis
implies different effects on risk aversion? Is the effect stronger or
weaker when one faces it once a grown-up?
Less restrictive (and more challenging) approach:
Does risk aversion changes over life? If so, what are the driving forces
of this change?
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 15 / 16

16. Conclusions Final words and debate
Final words and debate
Risk aversion mesurement
→ accurate? Should we really use it?
Arguable assumption
→ credible? What do you think?
Few control variables
→ add some? Suggestions?
Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 16 / 16