Childhood’s Economic Framworkand its Effects on Preferences

Melika Liporace

HEC Lausanne

April 17, 2014

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 1 / 16

Outline

1 Introduction

2 DataRisk AversionBirth Date

3 MethodologyAssumptionControl Variables

4 Preliminary resultsControl VariablesSignificance

5 ConclusionsExtensionsFinal words and debate

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 2 / 16

Introduction

Introduction

Main question:Does the economic framework of one’s childhood influences one’srisk aversion during later life?

Motivation:

HRS Data naturally offers a lot of information on effects of economicchanges through history

Risk Aversion is often ruled out as an heterogeneous feature, and notreally analyzed

Childhood’s Economic Framework can be seen almost as a randomvariable, reducing the risks of endogeneity often encountered inpreferences and microbehavioral analysis

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 3 / 16

Data Risk Aversion

Data: Risk AversionDesign

Two questions are asked:

Suppose that you are the only income earner in the family, and youhave a good job guaranteed to give you your current (family) incomeevery year for life. You are given the opportunity to take a new andequally 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 athird. 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 thenew 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 thentake the new job?

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 4 / 16

Data Risk Aversion

Data: Risk AversionDesign

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:I Two additionnal questions for Yes/Yes and No/No ⇒ two additional

categories ⇒ ”rescale”:

Wave 4 type 1 2 3 4 5 6Wave 1 type 1 1 2 3 4 4

I Question rephrased to avoid ”status-quo bias”

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 5 / 16

Data Risk Aversion

Data: Risk AversionSummary Stastics: relative frequencies

Relative Freq.

1 12.602 10.263 13.604 63.55Total 100.00

020

4060

Perc

ent

0 1 2 3 4Risk Aversion

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 6 / 16

Data Risk Aversion

Data: Risk AversionRelevance

Pros:I Measured by a specific design as well adapted as possibleI No utility specification assumedI Gamble over lifetime income (vs. usually, small bets)I Barsky et al. (1997)1 found that ”The measured risk tolerance has

predictive power for choices over risky behaviors - the decisions tosmoke and drink, to buy insurance, to immigrate, to be self-employed,and to hold stock”

Cons:I Self-reported choicesI Hypothetical situationsI Status-quo biasI Mitgated findings of Barsky et al.: ”the incremental predictive power of

risk tolerance is never very high.”

1Preference Paramters and Behavioral Heterogeneity: an Experimental Approach In Health and Retirment Study

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 7 / 16

Data Birth Date

Data: Birth DateSummary Statistics: Frequency

020

040

060

080

0Fr

eque

ncy

1900 1910 1920 1930 1940 1950 1960 1970 1980Birth Date

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 8 / 16

Data Birth Date

Data: Birth DateEconomic 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 0.1325World War II Dummy 5650 0.3468

Observations 16292

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 9 / 16

Data Birth Date

Data: Birth DateEconomic 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

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:I Latent variable riskavers∗ = ecodummies β + zγ + εI Outcome riskavers ∈ (0; 4), as the usual function of y∗

I Interested in partial effects of ecodummies (partially through β)I Control variables z

→ which control variables ?

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 11 / 16

Methodology Control Variables

Methodology: Control Variables

Idea : find variables that:I are correlated to individual’s birth dateI 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:I risk aversion could be partly inherited through upbringingI 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

Preliminary results Control Variables

Preliminary results: Control Variables

Mean

Years of Education 12.60Parent’s Education 10.33

Observations 14802

Control variables requirements :drops less than 2000

Risk aversion and Economicdummies distribution roughlystay the same

⇒ Not too demanding

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 13 / 16

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 issignificant

SurpriseI parent’s education very

significantI education not

significant at all

⇒ Should we collude bothdummies in only one suchas ”Before/after WWII”?

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 14 / 16

Conclusions Extensions

Extensions

Further analysis:

Does the age at which one has encountered an economic crisisimplies different effects on risk aversion? Is the effect stronger orweaker 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 forcesof this change?

Microeconometric Project Economic Framwork and Risk Aversion April 17, 2014 15 / 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