Lecture 24, Apr 18

Consumption, Saving, Interest Rates

Macroeconomics HUL213
A two-period model of consumption

1
A two-period model of consumption
• For now: study problem of individual in isolation taking as given prices
(partial equilibrium)
• In a bit: general equilibrium
• Two time periods t = 1 and t = 2
• Consumption c1 and c2 , income y1 and y2
• Utility function
u(c1 ) + βu(c2 )
with u strictly increasing, concave, discount factor 0 < β < 1
• Vocabulary that sometimes comes up: discount factor vs discount rate
• discount factor: β, typically a bit below one
• discount rate: ρ, typically a bit above zero
• link: ρ = 1/β − 1 ⇔ β = 1+ρ 1
, e.g. β = 0.95, ρ = 1/0.95 − 1 = 5.26% 2
A two-period model of consumption

Household solves
max u(c1 ) + βu(c2 ) s.t.
c1 ,c2 ,a
c1 + a = y1 (1)
c2 = y2 + (1 + r )a (2)
Notation:
• c1 , c2 : consumption at t = 1 and t = 2
• y1 , y2 : income at t = 1 and t = 2
• r : interest rate (for now exogenously given)
• a: saving
• Note: a can be negative, a < 0 means household is borrowing
• From (1) a < 0 ⇒ c1 > y1 , i.e. consume more than income by borrowing
3
Formulation in terms of present-value budget constraint

• Implicit assumption: can borrow and save as much you want at rate r
• then can combine (1) and (2) into present-value budget constraint
c1 y2 c2 y2
c1 +a = y1 and a = − ⇒ c1 + = y1 +
1+r 1+r | {z + r}
1 | {z + r}
1
PV of consumption PV of income

where PV stands for present value

• Hence households maximize utility s.t. present-value budget constraint

max u(c1 ) + βu(c2 ) s.t.

c1 ,c2
c2 y2
c1 + = y1 +
1+r 1+r
4
Euler equation

• Optimality condition

u ′ (c1 ) = β(1 + r )u ′ (c2 ) (∗)

• Intuition: consume $1 now vs save and consume later

• consume $1 now: utility increases by u ′ (c1 )
• save $1: have $(1 + r ) at t = 2, utility by βu ′ (c2 ) for each $

• (∗) is called “Euler equation” after Leonard Euler, plays an important

role in macroeconomics

• Note: “Euler equation” sounds fancy but simply means “intertemporal

optimality condition”
5
curve is tangent to the budget constraint. Notice two properties of the budget constraint. First, its slope is

− (1 + r). As usual, the slope of the budget constraint is the relative price. Higher interest rates mean a

Graphical representation
steeper budget constraint. Second, the budget constraint goes through the point (y1 , y2 ) since the household

has the option to just consume its income each period.

Fig. 6.2.1: The consumption-

savings decision as a two-good
consumption problem.

We can also nd the solution to problem (6.2.5) from its rst order conditions. The Lagrangian is:
5
6
A saver (left panel) and a borrower (right panel)

Left panel: y2 = 0 and y1 > 0 ⇒ save. Right panel: y1 << y2 ⇒ borrow.

7
What happens when r changes?

8
Effect of change in r for saver (left) and borrower (right)

9
The permanent income hypothesis

10
An important observation about our model
• Can write two-period model as
c2
max u(c1 ) + βu(c2 ) s.t. c1 + =W
c1 ,c2 1+r
y2
where W = y1 + = PV of income
1+r
sometimes also called “lifetime income” or “permanent income”
• Can immediately see: c1 will depend on y1 , y2 only through W , i.e.
y2
c1 = c(W ), W = y1 +
1+r
• It’s not current income that matters, but PV of current + future income
• This way of thinking: “permanent income hypothesis” (Friedman, 1957)
• Contrast with Keynes quote, captured w consumption function
c1 = c(y1 )
• Can already see: our model is really very different from Keynes view
11
A parametric example
1
c 1− σ − 1
u(c) = (∗∗)
1 − σ1
• σ is called “intertemporal elasticity of substitution (IES)”
• Sometimes 1/σ is called “coefficient of relative risk aversion (RRA)” and
(∗∗) is called “CRRA utility”
• Can show: log utility = special case with σ = 1 (use l’Hopital’s rule)
1
c 1− σ − 1
u(c) = lim = log c
σ→1 1 − 1
σ
• Note: weird −1 in numerator only there because we want to take this
limit
• Important for those reading Kurlat: he, many others use σ for coeff of
RRA
c 1−σ c 1−σ − 1
u(c) = or u(c) = 12
1−σ 1−σ
Euler equation in parametric example

• With functional form (∗∗) have

u ′ (c) = c −1/σ
• Euler equation
−1/σ −1/σ
u ′ (c1 ) = β(1 + r )u ′ (c2 ) ⇒ c1 = β(1 + r )c2
or
c2
= [β(1 + r )]σ
c1
• IES σ governs responsiveness of growth rate of consumption c2 /c1 to
changes in r and β, e.g.
∂ log(c2 /c1 )
= σ ⇒ hence the name IES
∂ log(1 + r )
• A low IES σ means households dislike intertemporal substitution (want
to smooth c) ⇒ low responsiveness of c2 /c1 to changes in r and β 13
Analytic solution in parametric example

• Not hard to show (you should be able to!): solution to problem is

 σ
1
β(1+r ) (1 + r ) 1+r
c1 =  σ W, c2 =  σ W
1 1
1 + β(1+r ) (1 + r ) 1 + β(1+r ) (1 + r )

where
y2
W = y1 +
1+r

• Useful special case (see next slide why): β(1 + r ) = 1

1+r
c1 = c2 = W
2+r
• If further r = 0, then c1 = c2 = W
2
14
Reasonable parameterizations

Reasonable parameterizations have two features

1. β(1 + r ) not too far from 1, equivalently discount rate ρ = 1/β − 1 ≈ r
• e.g. ρ = 0.05, r = 0.01 but not ρ = 0.5, r = 0.01 or ρ = 0.05, r = 0.5

• in general equilibrium model of part 4, 1 + r ∗ ≈ 1/β

• market discounts future at roughly the same rate as individuals

2. σ not too large

• e.g. σ = 1/2, probably ≤ 1, definitely not σ = 5 or 10
• empirical evidence: people do not massively substitute
intertemporally, i.e. they do not massively increase spending when
r↓
15
MPC out of transitory income shock
• Question: suppose y1 increases by $1 but y2 is unchanged. By how much
does c1 increase? What is MPC out of transitory income shock?

• Answer: the marginal propensity to consume (MPC) out of a transitory

income shock is  σ
1
∂c1 β(1+r ) (1 + r )
=  σ
∂y1 1+ 1
(1 + r )
β(1+r )
• Contrast with Keynesian cross: C = α + γ(Y − T ), γ = exogenously
given
• In current model, MPC is instead endogenous and depends on
preferences (β, σ) and prices (r )!
• This is precisely what we mean when we say “Keynesian cross is not
microfounded but modern macro models are”
16
MPC out of transitory income shock
• Question: suppose y1 increases by $1 but y2 is unchanged. By how much
does c1 increase? What is MPC out of transitory income shock?
• Answer: the marginal propensity to consume (MPC) out of a transitory
income shock is  σ
1
∂c1 β(1+r ) (1 + r )
=  σ
∂y1 1+ 1
(1 + r )
β(1+r )
• Also note: this MPC is a number < 1 and << 1 for reasonable
parameters
• In fact, in this 2-period model, MPC ≈ 1/2, e.g. with β(1 + r ) = 1
∂c1 1+r
=
∂y1 2+r
∂c1
and if r = 0 then ∂y1 = 1/2
17
MPC out of permanent income shock
• Question: suppose both y1 and y2 increase by $1. By how much does c1
increase? What is MPC out of permanent income shock?

• Assumption: y1 = y¯1 + ∆, y2 = ȳ2 + ∆ and ∆ increases

y2
• From W = y1 + 1+r : when both y1 , y2 ↑ by $1, W ↑ by ≈ $2

∂W 1 2+r
=1+ = ≈2
∂∆ 1+r 1+r
• Similarly  σ
1
∂c1 (2 + r )
β(1+r )
=  σ
∂∆ 1
1 + β(1+r (1 + r )
)

18
For reasonable parameters this is a number ≈ 1. Exact when β(1 + r ) = 1

∂c1
=1
∂∆
Main prediction of permanent income hypothesis (PIH):

MPC out of transitory income << MPC out of permanent income

19