Applied Economics

Decisions over time

Decisions over time 1 / 64

Decisions over time

Many decisions involve trade-offs between costs and benefits

occurring at different point in time
Decision and time
Atemporal decisions
Decision of goods to be consumed in current period
Intertemporal decisions
Decisions of goods to be consumed in current and future periods
Decisions of goods to be consumed in different future periods
Examples
Financial: Borrowing and saving
Health: Exercise, diet and lifestyle
Human capital: Education
Problem: Maximize utility when

Decisions over time 2 / 64

Road map

Decisions over time

Time consistency (neoclassical model)
Time inconsistency
Financial literacy

Decisions over time 3 / 64

Decisions over time

Decisions over time

Decisions over time 4 / 64

Decisions over time

Intertemporal choices
Individuals trade money and goods today for money and goods later.
The trade-off between now and later is called time discounting.
The rate of discounting expresses how much of future consumption and
money an individual is willing to give up for consumption and money
today.
The rate of time discounting vary between individuals.
Present and future values
A SEK today is worth more than a SEK tomorrow since you can do
something with it today, for example invest.

Decisions over time 5 / 64

Future value

If I put 10000 in my bank account today and interest rate is 10% per
year. How much will I have on my account in 1 year?

10000 × (1 + 0.1) = 11000

Future value
yT +t = yT × (1 + r )t
y = value
y = value
T = today
t = number of periods(e.g ., years)
r = discount rate
Present value = yT
Future value = yT +t

Decisions over time 6 / 64

Compound interest

Compound interest is the interest you earn on interest.

Future value after t periods

yT +t = yT × (1 + r )t

Future value after 3 periods

yT +t = yT × (1 + r )3 = yT × (1 + r ) × (1 + r ) × (1 + r )

An investment of 10000 SEK today with an annual interest 10% is

not 13000 SEK in 3 years but 13310 SEK.

y3 = 10000 × (1 + 0.1) × (1 + 0.1) × (1 + 0.1) = 13310

Decisions over time 7 / 64

Present value

Present value
yT +t
yT =
(1 + r )t
If I would like to have 10000 in my bank account in 1 year and
interest rate is 10%. How much should I put on the account today?
10000
yT =
(1 + 0.1)

yT = 9090

Decisions over time 8 / 64

Compound interest and pensions

You are offered a choice between three different saving plans for
retirement. They will all be paid to you when you retire at age of 68
with an annual interest rate of 10%. Which plan would you choose?
Would your choice change if annual interest rate was 3%?
Interest rate Plan 1 Plan 2 Plan 3
25 5000 0 25000
26-34 5000 per year 0 0
35-68 0 5000 per year 0
Total payment 50000 170000 25000

Decisions over time 9 / 64

Compound interest and pensions

Future value at age of 68

Age Plan 1 Plan 2 Plan 3
10% 2035820 1227383 1506001
3% 156591 288651 89113

If high interest rate (or yield) beneficial with high savings early in life
If low interest rate (or yield) beneficial with high total life savings

Decisions over time 10 / 64

Inflation

Inflation is a general increase in the prices of goods

Consumer price index is often used to measure inflation
Real interest rate measures the real cost of money to the borrower
and the real yield to the lender. The Fisher equation:

(1 + r n )
(1 + r ) =
(1 + r i )

where r is the real interest rate, r n is the nominal interest rate and r i
is the inflation rate
Nominal interest rate measures the interest rate when taking both
real interest rate and inflation into account. Nominal interest rate is
stated by banks.
For simplicity we assume inflation is equal to 0 unless otherwise
stated.
Decisions over time 11 / 64
Time consistency

Time consistency

Decisions over time 12 / 64

Two-period intertemporal model (Neoclassical model)

Time consistency refers to a situation where individuals’ preferences

do not change over time (that is plans do not change over time)
Intertemporal decisions in a two-period model
Period 1 - the current period
Period 2 - the future period
Trade-off between consuming today or in the future (trade-off
between consumption and saving)
By saving today, an individual gives up consumption today for
consumption in the future
By borrowing (negative savings) today, an individual consumes more
today but has to give up consumption in the future to repay the loan

Decisions over time 13 / 64

Income - Endowment

A static analyses of intertemporal choices (e.g., do not consider

choice of labour supply (and leisure))
An individual receives an exogenous income in both periods
y1 real income after tax in the current period
y2 real income after tax in the future period
c2

Endowment
y2

c1
y1

Decisions over time 14 / 64

Intertemporal budget constraint

Assume individuals’ can borrow and lend money with the same
interest rate r
An individual can exchange 1 unit of consumption in Period 1 for (1
+ r) units of consumption in Period 2
Consumption in Period 1

c 1 = y 1 − s1 ,

where s1 = saving in Period 1

If s1 > 0, the individual is saving - a saver
If s1 < 0, the individual is dissaving - a borrower

Decisions over time 15 / 64

Intertemporal budget constraint

Consumption in Period 2

c 2 = y 2 − s2 ,

where s2 = saving in Period 2

If s2 < 0, the individual is using savings from Period 1 (a saver in
Period 1)
If s2 > 0, the individual is paying back loan from Period 1 (a borrower
in Period 1)

Decisions over time 16 / 64

Intertemporal budget constraint

Maximum consumption in Period 1 (consumption in Period 1 if

nothing is consumed in Period 2) (compare Present Value)

1
c1 = y1 + y2
(1 + r )

Maximum consumption in Period 2 (consumption in Period 2 if

nothing is consumed in Period 1) (compare Future Value)

c2 = (1 + r )y1 + y2

Decisions over time 17 / 64

Two-periods and intertemporal budget constraint

c2

(1 + r )y1 + y2 Slope= −(1 + r )

Endowment(y1 , y2 )
y2

c1
y1 y + 1
1 (1+r ) y2

The slope of the budget constraint shows that opportunity cost of 1

unit of consumption in Period 1 corresponds to (1+r) units of
consumption in Period 2.

Decisions over time 18 / 64

Two-periods and intertemporal budget constraint
c2

(1 + r )y1 + y2 Slope= −(1 + r )

Endowment(y1 , y2 )
y2

c1
y1 y + 1
1 (1+r ) y2

An individual consuming less than y1 in Period 1 is a saver in Period

1 (and thus consuming more than y2 in Period 2)
An individual consuming more than y1 im Period 1 is borrower in
Period 1 (and thus consuming less than y2 in Period 2)

Decisions over time 19 / 64

Example - Two-periods and intertemporal budget
constraint
Income (endowment)
Income in current period: y1 = 1000
Income in future period y2 = 1500
Interest rate: r = 20%
Budget constraint
Maximum consumption in Period 1
1
c 1 = y1 + y2
(1 + r )
1
c1 = 1000 + 1500 = 2250
(1 + 0.2)
Maximum consumption in Period 2

c2 = (1 + r )y1 + y2

c2 = (1 + 0.2)1000 + 1500 = 2700

Decisions over time 20 / 64
Example - Two-periods and intertemporal budget
constraint

Budget line
c2 = 2700 − 1.2 ∗ c1
Different consumption choices
If all consumption in Period 2
c2 = 2700 − 1.2 ∗ 0 = 2700
If all consumption in Period 1
c2 = 2700 − 1.2 ∗ c1
0 = 2700 − 1.2 ∗ c1
c1 = 2700/1.2
c1 = 2250

Decisions over time 21 / 64

Example - Two-periods and intertemporal budget
constraint

c2

2700

1500

c1
1000 2250
An individual consuming less than 1000 i Period 1 is a saver
An individual consuming more than 1000 i Period 1 is a borrower

Decisions over time 22 / 64

Intertemporal indifference curves
In the atemporal case with 2 goods
The indifference curve shows combinations of 2 goods providing equal
utility to an individual
The indifference curves are assumed to be convex (”average of
consumption of both goods is better than extreme consumption of one
of the goods”)
The Marginal Rate of Substitution (MRS) is the rate at which a
consumer is willing to give up a very small amount of one good for
some of the other good in order to be attain exactly the same utility as
before the trade.
In the intertemporal case with 2 periods
The indifference curve shows combinations of consumption in the 2
periods providing equal utility to an individual
The indifference curves are assumed to be convex
The MRS is the rate at which a consumer is willing to give up a very
small amount of consumption in one period for some more
consumption in the other period in order to be attain exactly the same
utility as before the trade.
Decisions over time 23 / 64
Intertemporal indifference curves

c2

Increasing utility

I1 I2
c1
The slope of the indifference curve shows the MRS of consumption in
Period 1 for consumption in Period 2

Decisions over time 24 / 64

Individual optimization

An individual chooses a combination of consumption in Period 1 and

period 2 that maximises her utility

MRS = −(1 + r )

In words: The point where the rate at which an individual is willing to

trade consumption in Period 1 for consumption in Period 2 is equal to
the relative price of Period 1 consumption in terms of Period 2
consumption, which is expressed as (1 + r)

Decisions over time 25 / 64

Individual optimization

c2

c2max
MRS = −(1 + r )

I1 I2
c1
c1max

Decisions over time 26 / 64

Patience versus impatience

c2

c2max

c2P
y2
c2I I Patience

I Impatience
c1
c1P y1 c1I c1max

A patient individual consumes less in Period 1 and more in Period 2

compared to an impatient individual.

Decisions over time 27 / 64

Changes in income (increase)

Increase in period 1 Increase in period 2

c2
C2

0
y2
y2
y2

c1
y1 y 0 c1
1 y1

The budget constraint does not result in a parallel shift if r > 0 since
one unit more of consumption in period 1 corresponds to (1+r) units
of consumption in period 2.

Decisions over time 28 / 64

Example: Increase in income in Period 1

Increase in income in period 1 from 1000 to 1500

Period 1 = 1500
Period 2 = 1500
Interest rate is 20%
Budget constraint
Maximum consumption in Period 1 = 1500 + 1500/1.2 = 2750
Maximum consumption in Period 2 = 1500 ∗ 1.2 + 1500 = 3300
Budget line
Future consumption= 3300 − 1.2 ∗ c1

Decisions over time 29 / 64

Example: Increase in income in Period 1

c2
3300

2700

1500

c1
1000 1500 2250 2750

Decisions over time 30 / 64

Changes in interest rates

Decrease in interest rate Increase in interest rate

c2 c2

y2 y2

c1 c1
y1 y1

Decisions over time 31 / 64

Example: Increase in interest rate

Income
Period 1 = 1000
Period 2 = 1500
Interest rate is 50%
Budget constraint
Maximum consumption in Period 1 = 1000 + 1500/1.5 = 2000
Maximum consumption in Period 2 = 1000 ∗ 1.5 + 1500 = 3000
Budget line
Future consumption = 3000 − 1.5 ∗ c1

Decisions over time 32 / 64

Example: Increase in interest rate and budget constraint

c2
3000
2700

2250 c1
2000

Decisions over time 33 / 64

Income effect and substitution effect for goods

There are two effects of a price change of a good

Income effect - change in demand since purchasing power changed
Substitution effect - change in demand since relative prices of goods
changed
Example: If a good increases in price (2 goods)
The good is relatively more expensive than the other good. The
demand for the good decreases (and increases for the other goods
which are now relatively cheaper) (substitution effect)
The increase in price reduces purchasing power and this will reduce
demand (income effect)

Decisions over time 34 / 64

Example: Increase in interest rate - saver in period 1
c2
B1
B10
B0 B
C
A

U(c1 , c2 ) = I1
U(c1 , c2 ) = I0

B10 B1 B0 c1

Total effect: ”A to B”
Substitutioneffect in Period 1: ”A to C” that is a decrease
Incomeeffect in Period 1: ”C to B” that is an increase
Substitutioneffect in Period 2: ”A to C” that is a increase
Incomeeffect in Period 2: ”C to B” that is an increase
Decisions over time 35 / 64
How does a change in interest rate affect consumption?

The Total effect of a change can be decomposed into

Substitution effect (SE)
Income effect (IE)

Effect Period Increase Decrease

Saver Borrower Saver Borrower
SE Period 1 Decrease Decrease Increase Increase
Period 2 Increase Increase Decrease Decrease
IE Period 1 Increase Decrease Decrease Increase
Period 2 Increase Decrease Decrease Increase
Total Period 1 ? Decrease ? Increase
Period 2 Increase ? Decrease ?

Decisions over time 36 / 64

Extensions

Higher interest rate on borrowing than saving

c2

y2

c1
y1

Constraint on borrowing (liquidity constraint)

c2

y2

c1
y1

Decisions over time 37 / 64

Time inconsistency

Time inconsistency

Decisions over time 38 / 64

Experiment (Read and van Leeuwen, 1998)

Design
If you were deciding today, would you choose fruit or chocolate for
today?
If you were deciding today, would you choose fruit or chocolate for next
week?
Results
Now Next week
Chocolate 70% 26%
Fruit 30% 74%
This shows that there is a preference for immediate gratification

Decisions over time 39 / 64

Intertemporal decisions (McClure et al., 2007)

Exercise
Imagine that you are very thirsty. What would you choose in Choice A,
B and C? What do you think people on average would choose in
Choice A, B and C?
Choice A: Juice now OR twice amount of juice in 5 minutes.
Choice B: Juice in 20 minutes OR twice amount of juice in 25 minutes.
Choice C: Assume 20 minutes have passed...Juice now OR twice
amount of juice in 5 minutes.

Decisions over time 40 / 64

Experiment (McClure et al., 2007)

Experiment with extremely thirsty subjects

Experimental design
Choice A: Juice now or twice amount of juice in 5 minutes (60% of
subjects choose first option).
Choice B: Juice in 20 minutes or twice amount of juice in 25 minutes
(30% of subjects choose first option).
This experiment shows that there is a preference for immediate
gratification (present bias). Other examples include credit cards or
loans with very high interest rates.

Decisions over time 41 / 64

Revisit McClure - What happens after 20 min?

An individual chose ”25 minutes” in Choice B (and ”now” in Choice

A)
Assume we wait 20 minutes. Since 20 minutes have passed, Choice B
has now become a choice between ”now” and ”5 minutes” as in
Choice A.
If an individual chooses juice now rather than waiting 5 min (reversed
the choice she made 20 minutes ago!), then she is time inconsistent
(present biased).

Decisions over time 42 / 64

Intertemporal decisions (McClure et al., 2007)

Intertemporal neoclassical model

Choice A
U(C0 ) = u(c0 ) + δ 5 u(c5 )
Choice B
U(C0 ) = δ 20 u(c20 ) + δ 25 u(c25 )
Choice C
U(C20 ) = δ 20 u(c20 ) + δ 25 u(c25 )
If an individuals are time consistent they would always choose early or
late in all choices (utility independent, stationary instantaneous utility
δ05 δ025 25
δ20
and consumption Independence over time, and 1 = δ020
= 20
δ20
= δ 5 ).

Decisions over time 43 / 64

Time consistency

The standard neoclassical model assumes time consistency implying

an individual has the same preferences for future events independently
when the utility is evaluated
Exponential discount model

U(Ct ) = u(ct ) + δ 1 u(ct+1 ) + δ 2 u(ct+2 ) + ....δ T u(cT )

, where 0 ≤ δ ≤ 1. The higher δ the higher is the patience.

Assumptions in the exponential discount model (implicit)
Utility independence - no preferences for pattern (e.g., order of holiday
destinations)
Stationary instantaneous utility - same utility across time (e.g.,
emotional state does not influence e.g., hungry)
Consumption independence over time - no effect of prior or future
periods (e.g., addiction)

Decisions over time 44 / 64

Beta-delta model (Laibson, 1997)

Beta-delta model allows for time inconsistency

Model

U(Ct ) = u(ct ) + β(δ 1 u(ct+1 ) + δ 2 u(ct+2 ) + ....δ T u(cT ))

, where 0 ≤ δ ≤ 1 , 0 ≤ β ≤ 1
β= weight given to non-instantaneous rewards
The lower the δ the more impatient
The lower the β the more impulsive

Decisions over time 45 / 64

Beta-delta model (Laibson, 1997)

Revisit McClure et al. (2007) experiment

Choice A:
U(Cnow ) = u(1)
U(C5min ) = βδ 5 u(2)
Choice B:
U(C20min ) = βδ 20 u(1)
U(C25min ) = βδ 25 u(2)

Decisions over time 46 / 64

Beta-delta model (Laibson, 1997)

Assume that β = 0.4 and δ = 1 (given time span is at most only 25

min)
Choice A:
U(Cnow ) = u(1)
U(C5min ) = 0.4u(2)
Choice B:
U(C20min ) = 0.4u(1)
U(C25min ) = 0.4u(2)
If we assume that one unit of juice equals one util
Choice A: Chooses ”Now” 1>0.8
Choice B: Chooses ”25 min” 0.4<0.8
The beta-delta model can explain the choices

Decisions over time 47 / 64

Revisit McClure et al. (2007)
Choice B: Juice in 20 minutes or twice amount of juice in 25 minutes
U(C20min ) = βδ 20 u(1)
U(C25min ) = βδ 25 u(2)
If we assume that one unit of juice equals one util, β = 0.4 and δ = 1
Choice B: Chooses ”25 min” 0.4<0.8
When 20 minutes have passed, then Choice C, will become like
Choice A (now or in 5 min):
U(Cnow ) = u(1)
U(C5min ) = 0.4u(2)
Choice A: Chooses ”Now” 1>0.8, which contradicts the Choice C
made 20 minutes ago!
If subjects were time consistent they would always choose early or late
in both choices, but we often observe intra-person conflicts.
Decisions over time 48 / 64
Naive and sophisticated individuals

Expectations about future time preferences at time (t + k)

Û(Ct ) = u(ct+k )+ β̂(δ 1 u(ct+k+1 )+δ 2 u(ct+k+2 )+....+δ T u(ct+k+T ))

, where β̂ is expectation of future weight given to non-instantaneous

rewards
Fully naive β < β̂ = 1
An individual behaves as if future plans will be followed (see previous
example).
Sophisticated β̂ = β < 1
An individual correctly expects future behaviour (expects to behave as
in Choice A when 20 minutes have passed).
Partially naive β < β̂

Decisions over time 49 / 64

Present biased preferences

Naive
The individuals are not aware that their future selves will be
present-biased, and hence will procrastinate for example work and
pension savings (or consume juice immediately).
Sophisticated
The individuals are aware that their future selves will be present-biased
and look for commitment devices. A commitment device is a tool that
either restricts an individuals future choices or make them more
expensive, but involve transaction costs and experiences of using and
uncertainties about effect.
blocking web pages or internet access while working
banning access to online games
www.stickk.com - Objective “stickK empowers you to better your
lifestyle. We offer you the opportunity, through ’Commitment
Contracts’, to show to yourself and others the value you put on
achieving your goals.”

Decisions over time 50 / 64

The marshmallow experiment

The marshmallow experiment was conducted by Walter Mischel at

Stanford University in 1972.
The objective was to investigate deferred gratification among
children, and if the success in resisting the temptation of eating the
marshmallow is correlated with future success.
More than 600 children aged 4-6 participated
Experimental design
The experiment took place in a room which was empty of distractions.
A marshmallow was placed on the desk in front of them. The children
could either eat the marshmallow or wait for fifteen minutes to get a
second marshmallow.

Decisions over time 51 / 64

The marshmallow experiment

A small minority ate the marshmallow immediately, and one third

managed to resist in order to get a second marshmallow. The ability
to resist was positively correlated by age.
Follow-up studies. . . .
In 1988, a study showed that ”preschool children who delayed
gratification longer in the self-imposed delay paradigm, were described
more than 10 years later by their parents as adolescents who were
significantly more competent”.
In 1990, it was found that the ability to delay gratification also
correlated with higher SAT scores.
In 2011, a neuroscience study showed significant differences in brain
images between those who deferred gratification (more activity in
prefrontal cortex) while the other showed a higher activity in ventral
striatum (area linked to addiction).

Decisions over time 52 / 64

Financial literacy

Financial literacy

Decisions over time 53 / 64

Financial literacy

Financial literacy - An individual has skills and knowledge to make

informed financial decisions
People make many financial decisions themselves
Savings and loans (both short-term (e.g., credit card) and long-term
(e.g., mortgages))
Investments in financial products (e.g., stocks, mutual funds and
options)
Pensions

Decisions over time 54 / 64

Financial inclusion

Financial inclusion - Access to financial institutions and financial

products such as bank accounts, credit cards, debit cards and loans.
Percentage of adults owning a bank account in each country

Source: Grohmann and Menkhoff (2017, p.400)

Decisions over time 55 / 64

Financial literacy - Survey questions

Numeracy (compounding of interest rates)

”Suppose you had 100 USD in a savings account and the interest rate
was 2 percent per year. After 5 years, how much do you think you
would have in the account if you left the money to grow:” [more than
102 USD; exactly 102 USD; less than 102 USD; do not know; refuse
to answer.] [Answer: more than 102 USD]
Inflation
”Imagine that the interest rate on your savings account was 1 percent
per year and inflation was 2 percent per year. After 1 year, would you
be able to buy:” [more than, exactly the same as, or less than today
with the money in this account; do not know; refuse to answer.]
[Answer: less than today]

Decisions over time 56 / 64

Financial literacy - Survey questions

Knowledge about “stocks” and “stock mutual funds” and of

diversification
”Do you think that the following statement is true or false? “Buying
a single company stock usually provides a safer return than a stock
mutual fund.” [true; false; do not know; refuse to answer.] [Answer:
false]

Decisions over time 57 / 64

Financial literacy - Understanding

Source: Lusardi and Mitchell (2014, p.14)

Decisions over time 58 / 64

Financial literacy - Understanding

Source: Lusardi and Mitchell (2014, p.18)

Source: Lusardi and Mitchell (2014, p.18)

Decisions over time 59 / 64

Financial literacy - Understanding

Source: Lusardi and Mitchell (2014, p.18)

Source: Lusardi and Mitchell (2014, p.19)

Decisions over time 60 / 64
Financial literacy and economic behavior
(Lusardi and Mitchell, 2014)

Difficult to establish a causal link between financial literacy and

economic behavior
Summary of some empirical findings
Financial literacy and participation in financial markets are positively
correlated
Financial literacy and retirement planning are positively correlated
Financial literacy and relatively lower interest rates mortgages are
positively correlated
Financial literacy and good credit card behaviour are positively
correlated

Decisions over time 61 / 64

Assessments of the effect of financial literacy programs
(Lusardi and Mitchell, 2014)

Challenge: The effect of a short exposure to a financial literacy

program (for examples a few meetings on pension savings) is unlikely
to affect economic behaviour
The quality of the financial literacy program (material and teachers)
is key for having the potential of being successful, but it is also
important that the program is tailor-made for the target population
Evaluation of financial literacy programs should be done using
experiments (with control and treatments) focusing on both short-
and long-term effects

Decisions over time 62 / 64

Assessments of the effect of financial literacy programs
(Lusardi and Mitchell, 2014)

Financial education programs have been offered to high school

students in the US (have been mandatory in several U.S. states)
It is also important to make financial decision-making easier for
individuals, for example the choice architecture.
Note that choice architects (e.g., financial institutions) might
maximize their own profit by using for example defaults favourable for
themselves

Decisions over time 63 / 64

References

Grohmann, A. and L. Menkhoff (2017). Financial literacy promotes

financial inclusion in both poor and rich countries, DIW Economic
Bulletin, 17.
Laibson, D. (1997). Golden eggs and hyperbolic discounting,
Quarterly Journal of Economics, 112, 443-477.
Lusardi, A. and O.S. Mitchell (2014). The Economic Importance of
Financial Literacy: Theory and Evidence, Journal of Economic
Literature, 52, 5–44.
McClure, S.M., K.M. Ericson, D. I. Laibson, G. Loewenstein and J.D.
Cohen (2007). Time Discounting for Primary The Journal of
Neuroscience, 27, 5796 –5804.
Read, D. and B. van Leeuwen (1998). Predicting hunger: The effects
of appetite and delay on choice, Organizational Behavior and Human
Decision Processes, 76, 189-205.

Decisions over time 64 / 64