savings

lecture 6

intertemporal risk

risk-coping over time: consumption smoothing = to average of income

risk-coping across people: cross-sectional insurance (informal insurance)

epsiloni = idiosyncratic risk

theta= aggregate risk

yi= A+ epsiloni + theta insurance

saving: yi= A + epsilon + delta

epsilon= transitory shock (sick one day + fluctuates overtime)

delta= permanent shock (permanently disabled)

negative transitory shock —> savings decrease

what gives variation that is not fixed overtime?

permanent income hypothesis

households choose each period’s consumption level based on (expected) permanent income

perfect foresight (know of shocks) —> set consumption every period + equal to average income

already have concavity in utility —> need smooth consumption

household’s goal: smooth over income

transitory shocks will not affect consumption but should be smoothed through savings(permanent income not change)

permanent shocks will not affect savings but affect consumption: one for one affect on consumption

paxson in thailand permanent income

savings function of transitory shock are not permanent

consumption function of permanent shock not transitory

info:

income-expenditures

income expenditures on everything but durables

purchases-sales of financial assets

how do savings rates vary in sensible ways?

most households have negative savings rates: when surveyed, moment of dissaving

variation in savings among households

paxson regressions

sirt (savings of household/individual)= alpha0 + alpha1yirt^p + alpha2yirt^t + epsilonirt

alpha 0 should be 0, alpha 1 should be 1

yirt^p= permanent income

yirt^t= transitory income

problem: how do we split permanent and transitory

empirical strategy

how do we identify different types of income?

estimates permanent income by looking at incomes across people

characteristics that don’t change much (permanent things)—> predict variation in income

-education

-demographics

-assets

estimate transitory income:

-rainfall data= transitory shock

xirt^p= household specific variables

testing pih

-test whether transitory rainfall shocks/permanent income affect consumption/savings. if they do, pih is rejected

-estimate how transitory rainfall shocks and permanent income affect income, and they see how much they affect consumption/savings

results:

r1-r1^- : difference b/w average rainfall and current. more rainfall than usual > 0

f-tests:

rainfall variables are jointly insignificant + reject that they don’t matter (p-value); double-negative= positive

rainfall predicts savings

permanent income has an affect on savings: reject pih (test 1 extensive) + but small effect

fixed characteristics= strong predictors of income

test 2 intensive: how correlated is income w/ savings

alpha 2 =1 : fail to reject

paxson summary

how much permanent + transitory change with rainfall and household variables —> then see how this affects consumption

2-step procedure:

can reject that the coefficient on permanent income=0

cannot reject that the coefficient on transitory income = 1

pih rejected but pretty close

economics of roscas

rosca= rotating savings + credit association

major method for savings for many very poor people

move money across different people = people bring a specified amount of money to ROSCA pot at each meeting

payouts can be random, bidding (auction to decide who gets the money), pre-determined (order decided beforehand)

value of bidding ROSCA

why ROSCA exist

financing durable good purchases

insurance/cash liquidity

“you can’t save alone”

protect female income= shield income from spouses