Outline

Loss Aversion, Education, and Inter-Generational Income Mobility

1. Introduction

a. Fair playing field

b. Number of factors contributing to inequality

i) Education

ii) Innate ability

iii) Starting point

(1) Intergenerational mobility (or lack thereof)

c. How do parents contribute?

i) Financially

(1) Allows children to reduce losses in first (education) period

ii) Genetically

(1) IQ

(2) Other factors

iii) Setting reference level of consumption

(1) Level at which children will shoot when making education decisions

iv) Borrowing constraints

(1) Limits possible education for children from lower end of income distribution if related to parent income

(2) Increases loss in first (education) period for higher levels of education

d. Importance of loss aversion

i) Children from higher income families will take a loss in period one in order to reach reference consumption in period two

2. Model

a. Setup

i) Overlapping generations (3 periods)

ii) “Earnings ability” iid

iii) Education choice/wage determination

iv) Loss aversion (reference consumption)

v) Bequest motive (warm glow)

vi) Borrowing constraint

b. Loss Aversion

i) Weak vs. Strong

ii) Consumption/Savings patterns

c. Borrowing constraints

i) Weak (borrow up to cost of education)

ii) Strong (borrow up to some fraction of family income)

d. Bequest motive

3. Review of the data

a. Income distribution

i) By education level

b. Intergenerational correlations

i) Income

ii) Education

c. Educational attainment by family income

i) Mixed once controlling for parental education

d. Transition matrix for income quantiles

4. Simulations

a. Parameter values

b. Variables

c. Main results

i) Percent at each education level

ii) Distribution of income

iii) Correlation of inter-generational income

iv) Correlation of inter-generational education

v) Correlation of inter-generational consumption

vi) Correlation of inter-generational bequests

vii) Transition matrix for simulated income quantiles

d. Results without loss aversion, borrowing constraints, bequests

i) No habit formation

ii) Habit formation

iii) Add loss aversion

iv) Add borrowing constraints

v) Add bequests

vi) Show how much each adds individually and how they work together

5. Conclusion

a. Model shows three main ways that parent income can be transferred to child income

i) Loss aversion

ii) Borrowing constraints (when strong)

iii) Bequests

b. Work together to explain a significant portion of the inter-generational correlation of income

i) But not the only source

c. Potential policy implications

i) While these results may be individually utility maximizing, they are not socially optimal

(1) Would like people to maximize their lifetime resources (max GDP)

ii) Possible ways to do that:

(1) Make education financing available to all

(a) Include consumption allowance

(2) Inheritance tax

Loss Aversion and Human Capital

My current project is on loss aversion, education choice, and inter-generational income mobility. I will try to write the introduction tomorrow.

The basic idea is that people develop a reference level of consumption in their first period of life (known as childhood to non-economists). They then must make an education choice in the next period, facing both a direct cost and an opportunity cost.

Because agents are loss averse, they may choose a lower level of education than would maximize lifetime resources in order to avoid losses in the education period. In the final period they earn a wage that is determined by both their education level and their "earnings ability."

There are a number of questions that still need to be answered in the paper:

1. Is loss aversion necessary, or can I just use habit formation with a standard concave utility?
2. Can parents take into account the utility of children in making their decisions, or is it ok if they just have "warm-glow" altruism?
3. Should the direct cost of education be convex or should the opportunity cost be convex? Which is easier to model?
4. Can the model successfully replicate education choice, inter-generational correlation of income, and the overall income distribution, or is that asking too much?
5. If wages are stochastic, what is the effect on the education decision? With loss aversion we won't have certainty equivalence, right?
6. What is the result of using strong vs. weak loss aversion? What are the consumption implications of weak loss aversion?
   
7. Should utility depend both on the absolute level of consumption and the difference between current consumption and the reference level? Would that help or hurt the results?
8. How important is each aspect of the model? Loss aversion, habit formation, borrowing constraint, and bequests.

It will be fun to find out. Maybe.