Repaso de Macroeconomía Avanzada

Unidad I: Modelos de generaciones traslapadas (OLG)

• McCandless, George T. and Wallace, Neil (1991). Introduction to dynamic macroeconomic

theory: an overlapping generations approach. Harvard University Press.
• Romer, David (2012). Advanced Macroeconomics. Fourth edition. McGraw-Hill. Chapter 2.

 Ganancias agregadas e intergeneracionales del intercambio

 Ineficiencia dinámica
 Gobierno, dinero y herencias en el modelo OLG

1 Infinite-horizon and overlapping-generations models

This chapter investigates two models that resemble the Solow model but in which the dynamics of
economic aggregates are determined by decisions at the microeconomic level.

Both models continue to take the growth rates of labor and knowledge as given. But the models
derive the evolution of the capital stock from the interaction of maximizing households and firms in
competitive markets.

Ramsey (1928), Cass (1965) and Koopmans (1965) Model:

Competitive firms rent capital and hire labor to produce and sell output, and a fixed number of
infinitely lived households supply labor, hold capital, consume, and save. Avoids all market
imperfections and all issues raised by heterogeneous households and links among generations. It
therefore provides a natural benchmark case

Diamond (1965) Model:

Is the overlapping-generations model assumes continual entry of new households into the
economy. As we will see, this seemingly small difference has important consequences.

1.1 The Ramsey–Cass–Koopmans Model

1.1.1 Assumptions
a. There are a large number of identical firms and each of these has access to the production
function 𝑌 = 𝐹(𝐾, 𝐴𝐿).
b. Firms take 𝐴 as given; as in the Solow model, 𝐴 grows exogenously at rate 𝑔
c. The firms maximize profits, are owned by the households and taking decitions in
competitive factor markets, and sell their output in a competitive output market.
d. The household divides its income (from the labor and capital it supplies and, potentially,
from the profits it receives from firms) at each point in time between consumption and
saving so as to maximize its lifetime utility.
 ∞
𝐿(𝑡)
𝑈 = ∫ 𝑒 −𝜌𝑡 𝑢(𝐶(𝑡)) 𝑑𝑡
𝑡=0 𝐻

𝐶(𝑡) is the consumption of each member of the household at time 𝑡.

𝑢(•) is the instantaneous utility function, which gives each member’s utility at a given date.

𝐿(𝑡) is the total population of the economy.

𝐿(𝑡)
𝐻
is therefore the number of members of the household.

𝐿(𝑡)
𝑢(𝐶(𝑡)) 𝐻
is the household’s total instantaneous utility at 𝑡.

𝜌 is the discount rate; the greater is 𝜌, the less the household values future consumption relative to
current consumption.

The instantaneous utility function takes the form of as constant-relative-risk-aversion (or CRRA):

𝐶(𝑡)1−𝜃
𝑢(𝐶(𝑡)) = , 𝜃 > 0, 𝜌 − 𝑛 − (1 − 𝜃) > 0
1−𝜃
𝑢′′(𝐶(𝑡))
−𝐶(𝑡) = 𝜃
𝑢′(𝐶(𝑡))

𝜃 determines the household’s willingness to shift consumption between different periods

1.1.2 The Behavior of Households and Firms

At each point in time they employ the stocks of labor and capital, pay them their marginal products,
and sell the resulting output. Because the production function has constant returns and the
economy is competitive, firms earn zero profits. As the markets are competitive, capital earns its
marginal product and there is no depreciation, the real rate of return on capital equals its earnings
per unit time.

the real interest rate at time t is:

𝑟 (𝑡 ) = 𝑓′(𝑘 (𝑡 ))
Taking account that:
𝐾
𝜕𝑓(𝐾, 𝐴𝐿) 𝜕𝐴𝐿𝑓 (𝐴𝐿) 𝜕𝐴𝐿𝑓(𝑘) 𝑓′(𝑘)
= = = 𝐴 [𝑓(𝑘) − 𝐿𝐾 ] = 𝐴[𝑓(𝑘) − 𝑘𝑓′(𝑘)] = 𝑊
𝜕𝐿 𝜕𝐿 𝜕𝐿 𝐴𝐿2

Formally with respetc to the time t:

𝑊(𝑡) = 𝐴(𝑡)[𝑓(𝑘(𝑡)) − 𝑘(𝑡)𝑓′(𝑘(𝑡))]

The wage per unit of effective labor is therefore:

𝑊(𝑡)
= 𝑤(𝑡) = 𝑓(𝑘(𝑡)) − 𝑘(𝑡)𝑓′(𝑘(𝑡))
𝐴(𝑡)
1.1.2.1 Households’ Budget Constraint

a. Taking the paths of 𝑟 and 𝑤 as given, the households’ Budget constraint is that the present
value of its lifetime consumption cannot exceed its initial wealth plus the present value of
its lifetime labor income.
b. Given taht 𝑟 may vary over time:
𝑡
𝑅(𝑡) = ∫ 𝑟(𝜏) 𝑑𝜏
𝜏=0

c. The household’s Budget constraint is therefore:

∞ ∞
𝐿(𝑡) 𝐾(0) 𝐿(𝑡)
∫ 𝑒 −𝑅(𝑡) 𝐶(𝑡) 𝑑𝑡 ≤ + ∫ 𝑒 −𝑅(𝑡) 𝑊(𝑡) 𝑑𝑡
𝑡=0 𝐻 𝐻 𝑡=0 𝐻
∞
𝐾(0) 𝐿(𝑡)
+ ∫ 𝑒 −𝑅(𝑡) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡 ≥ 0
𝐻 𝑡=0 𝐻
𝐾(0)
d. Where initial capital holdings of 𝐻
, where 𝐾(0) is the initial amount of capital in the
economy and 𝐻 is the number of households.
𝐿(𝑡) 𝐿(𝑡)
e. The household has 𝐻
members, its labor income at 𝑡 is 𝑊(𝑡) 𝐻 , and its consumption
𝐿(𝑡) 1 𝐾(0)
expenditures are 𝐶(𝑡) 𝐻 . Its initial wealth is 𝐻 of total wealth at time 0, or 𝐻

We can write the integral from 𝑡 = 0 to 𝑡 = ∞ as a limit:

𝑆
𝐾(0) 𝐿(𝑡)
lim { + ∫ 𝑒 −𝑅(𝑡) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡} ≥ 0
𝑆→∞ 𝐻 𝑡=0 𝐻
𝑆 𝑆
𝐾(𝑆) 𝐿(𝑡) 𝐾(0) 𝑅(𝑆) 𝐿(𝑡)
− ∫ 𝑒 𝑅(𝑆) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡 − [ 𝑒 + ∫ 𝑒 −𝑅(𝑡) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡] ≥ 0
𝐻 𝑡=0 𝐻 𝐻 𝑡=0 𝐻
𝑆
𝐾(𝑆) 𝐾(0) 𝑅(𝑆) 𝐿(𝑡)
≥ 𝑒 + ∫ 𝑒 𝑅(𝑆)−𝑅(𝑡) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡
𝐻 𝐻 𝑡=0 𝐻
𝑆
𝐾(𝑆) −𝑅(𝑆) 𝐾(0) 𝐿(𝑡) 𝐾(𝑆) −𝑅(𝑆)
𝑒 ≥ + ∫ 𝑒 𝑅(𝑆)−𝑅(𝑡) [𝑊(𝑡) − 𝐶(𝑡)] 𝑑𝑡 → lim [ 𝑒 ]≥0
𝐻 𝐻 𝑡=0 𝐻 𝑆→∞ 𝐻
𝐾(0) 𝑅(𝑆)
𝐻
𝑒 is the contribution of the household’s initial wealth to its wealth at 𝑠.
𝐿(𝑡)
The household’s saving at 𝑡 is [𝑊(𝑡) − 𝐶(𝑡)] 𝐻
(which may be negative).

𝑒 𝑅(𝑆)−𝑅(𝑡) shows how the value of that saving changes from 𝑡 to 𝑠.

The budget constraint states that the present value of the household’s asset holdings cannot be
negative in the limit (is known as the no-Ponzi-game condition).
1.1.2.2 Households’ Maximization Problem
The representative household wants to maximize its lifetime utility subject to its budget constraint.
Now let´s define 𝑐(𝑡) to be consumption per unit of effective labor, where 𝐶(𝑡) = 𝐴(𝑡)𝑐(𝑡). The
household’s instantaneous utility is:

𝐶(𝑡)1−𝜃 [𝐴(𝑡)𝑐(𝑡)]1−𝜃 [𝐴(𝑡)]1−𝜃 [𝑐(𝑡)]1−𝜃

= = → 𝐴(𝑡) = 𝐴(0)𝑒 𝑔𝑡
1−𝜃 1−𝜃 1−𝜃
𝐶(𝑡)1−𝜃 [𝐴(0)𝑒 𝑔𝑡 ]1−𝜃 [𝑐(𝑡)]1−𝜃 𝑐(𝑡)1−𝜃 𝑔𝑡(1−𝜃)
= = 𝑒 𝐴(0)1−𝜃
1−𝜃 1−𝜃 1−𝜃
Substituting the fact that 𝐿(𝑡) = 𝐿(0)𝑒 𝑛𝑡 :
∞ ∞
𝐿(𝑡) 𝐶(𝑡)1−𝜃 𝐿(0)𝑒 𝑛𝑡
𝑈 = ∫ 𝑒 −𝜌𝑡 𝑢(𝐶(𝑡)) 𝑑𝑡 = ∫ 𝑒 −𝜌𝑡 𝑑𝑡
𝑡=0 𝐻 𝑡=0 1−𝜃 𝐻
∞
𝑐(𝑡)1−𝜃 𝑔𝑡(1−𝜃) 𝐿(0)𝑒 𝑛𝑡
𝑈 = ∫ 𝑒 −𝜌𝑡 𝑒 𝐴(0)1−𝜃 𝑑𝑡
𝑡=0 1−𝜃 𝐻

𝐿(0) ∞ [𝑔(1−𝜃)+𝑛−𝜌]𝑡 𝑐(𝑡)1−𝜃

𝑈 = 𝐴(0)1−𝜃 ∫ 𝑒 𝑑𝑡
𝐻 𝑡=0 1−𝜃
∞
𝑐(𝑡)1−𝜃
𝑈 = 𝐵 ∫ 𝑒 −𝛽𝑡 𝑑𝑡
𝑡=0 1−𝜃
𝐿(0)
𝐵 = 𝐴(0)1−𝜃 , 𝛽 = 𝑔(1 − 𝜃) + 𝑛 − 𝜌
𝐻
Thus we can rewrite:
∞ ∞
𝐿(𝑡) 𝐾(0) 𝐿(𝑡)
∫ 𝑒 −𝑅(𝑡) 𝐶(𝑡) 𝑑𝑡 ≤ + ∫ 𝑒 −𝑅(𝑡) 𝑊(𝑡) 𝑑𝑡
𝑡=0 𝐻 𝐻 𝑡=0 𝐻

If 𝐶(𝑡) = 𝐴(𝑡)𝑐(𝑡), 𝐾(0) = 𝐴(0)𝑘(0)𝐿(0), 𝑊(𝑡) = 𝐴(𝑡)𝑤(𝑡).

∞ ∞
𝐴(𝑡)𝐿(𝑡) 𝐴(0)𝐿(0) 𝐴(𝑡)𝐿(𝑡)
∫ 𝑒 −𝑅(𝑡) 𝑐(𝑡) 𝑑𝑡 ≤ 𝑘(0) + ∫ 𝑒 −𝑅(𝑡) 𝑤(𝑡) 𝑑𝑡
𝑡=0 𝐻 𝐻 𝑡=0 𝐻
∞
𝐴(0)𝑒 𝑔𝑡 𝐿(0)𝑒 𝑛𝑡 𝐴(0)𝐿(0) ∞
𝐴(0)𝑒 𝑔𝑡 𝐿(0)𝑒 𝑛𝑡
∫ 𝑒 −𝑅(𝑡) 𝑐(𝑡) 𝑑𝑡 ≤ 𝑘(0) + ∫ 𝑒 −𝑅(𝑡) 𝑤(𝑡) 𝑑𝑡
𝑡=0 𝐻 𝐻 𝑡=0 𝐻
∞ ∞
∫ 𝑒 −𝑅(𝑡) 𝑐(𝑡)𝑒 (𝑔+𝑛)𝑡 𝑑𝑡 ≤ 𝑘(0) + ∫ 𝑒 −𝑅(𝑡) 𝑤(𝑡)𝑒 (𝑔+𝑛)𝑡 𝑑𝑡
𝑡=0 𝑡=0

Finally, because 𝐾(𝑠) is proportional to 𝑘(𝑠)𝑒 (𝑔+𝑛)𝑆 , we can rewrite the no-Ponzi-game version of
the budget constraint:

lim [𝑒 −𝑅(𝑡) 𝑘(𝑠)𝑒 (𝑔+𝑛)𝑆 ] ≥ 0

𝑆→∞
1.1.2.3 Household Behavior
The household’s problem is to choose the path of 𝑐(𝑡) to maximize lifetime utility subject to the
budget constraint:
∞
𝑐(𝑡)1−𝜃 ∞ ∞
ℒ = 𝐵 ∫ 𝑒 −𝛽𝑡 𝑑𝑡 + 𝜆 {𝑘(0) + ∫ 𝑒 −𝑅(𝑡) 𝑤(𝑡)𝑒 (𝑔+𝑛)𝑡 𝑑𝑡 − ∫ 𝑒 −𝑅(𝑡) 𝑐(𝑡)𝑒 (𝑔+𝑛)𝑡 𝑑𝑡}
𝑡=0 1−𝜃 𝑡=0 𝑡=0

The household chooses 𝑐 at each point in time; that is, it chooses infinitely many 𝑐(𝑡)’s. The first-
order condition for an individual 𝑐(𝑡) is:

𝑐(𝑡): 𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 − 𝜆𝑒 −𝑅(𝑡) 𝑒 (𝑔+𝑛)𝑡 = 0

For the behavior of consumption, first take logs of both sides:

ln 𝐵 − 𝛽𝑡 − 𝜃 ln 𝑐(𝑡) = ln 𝜆 − 𝑅(𝑡) + (𝑔 + 𝑛)𝑡

𝑡
ln 𝐵 − 𝛽𝑡 − 𝜃 ln 𝑐(𝑡) = ln 𝜆 − ∫ 𝑟(𝜏) 𝑑𝜏 + (𝑔 + 𝑛)𝑡
𝜏=0

The derivatives of the two sides with respect to 𝑡 must be:

̇
𝑐(𝑡)
−𝛽 − 𝜃 = −𝑟(𝑡) + (𝑔 + 𝑛)
𝑐(𝑡)
If 𝛽 = 𝜌 − 𝑔(1 − 𝜃) − 𝑛 and replacing:
̇
𝑐(𝑡) 𝑟(𝑡) − 𝑔 − 𝑛 − 𝛽 𝑟(𝑡) − 𝑔 − 𝑛 − (𝜌 − 𝑔(1 − 𝜃) − 𝑛) 𝑟(𝑡) − 𝜌 − 𝑔𝜃 𝑟(𝑡) − 𝜌
= = = = −𝑔
𝑐(𝑡) 𝜃 𝜃 𝜃 𝜃
So the Euler equation for this maximization problema is:
̇
𝑐(𝑡) 𝑟(𝑡) − 𝜌
= −𝑔
𝑐(𝑡) 𝜃
We can denote that:

𝐶(𝑡) = 𝐴(𝑡)𝑐(𝑡)
ln 𝐶(𝑡) = ln 𝐴(𝑡) + ln 𝑐(𝑡)
̇
𝐶(𝑡) ̇
𝐴(𝑡) ̇
𝑐(𝑡) 𝑟(𝑡) − 𝜌 𝑟(𝑡) − 𝜌
= + =𝑔+ −𝑔 =
𝐶(𝑡) 𝐴(𝑡) 𝑐(𝑡) 𝜃 𝜃
This condition states that consumption per worker is rising if the real return exceeds the rate at
which the household discounts future consumption, and is falling if the reverse holds. The smaller
is 𝜃 —the less marginal utility changes as consumption changes— the larger are the changes in
consumption in response to differences between the real interest rate and the discount rate.

If the household reducing 𝑐 at some date 𝑡 by a infinitesimal amount ∆𝑐, investing this additional
saving for a infinitesimal period of time ∆𝑡, and then consuming the proceeds at time 𝑡 + ∆𝑡; assume
that when it does this, the household leaves consumption and capital holdings at all times other
than 𝑡 and 𝑡 + ∆𝑡 unchanged. If the household is optimizing, the marginal impact of this change on
lifetime utility must be zero. If the impact is strictly positive, the household can marginally raise its
lifetime utility by making the change. And if the impact is strictly negative, the household can raise
its lifetime utility by making the opposite change.

the marginal utility of 𝑐(𝑡) is:

𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃
Thus the change has a utility cost of:

𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐

Since the instantaneous rate of return is 𝑟(𝑡), 𝑐 at time 𝑡 + ∆𝑡 can be increased by 𝑒 [𝑟(𝑡)−𝑔−𝑛 ]∆𝑡 ∆𝑐,
this implies that:
̇
𝑐(𝑡)
[𝑐(𝑡) ]∆𝑡
[𝑟(𝑡)−𝑔−𝑛 ]∆𝑡
𝑐(𝑡 + ∆𝑡) = 𝑐(𝑡)𝑒 ∆𝑐 = 𝑐(𝑡)𝑒 ∆𝑐
Thus the marginal utility of 𝑐(𝑡 + ∆𝑡 ) is:

̇ −𝜃
𝑐(𝑡)
[𝑐(𝑡) ]∆𝑡
𝑈𝑀𝑔𝐶 = 𝐵𝑒 −𝛽[𝑡+∆𝑡] 𝑐(𝑡 + ∆𝑡)−𝜃 = 𝐵𝑒 −𝛽[𝑡+∆𝑡] [𝑐(𝑡)𝑒 ∆𝑐]

For the path of consumption to be utility-maximizing, it must therefore satisfy:

̇ −𝜃
𝑐(𝑡)
[𝑐(𝑡) ]∆𝑡
𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐 = 𝐵𝑒 −𝛽[𝑡+∆𝑡] [𝑐(𝑡)𝑒 ∆𝑐] 𝑒 [𝑟(𝑡)−𝑔−𝑛 ]∆𝑡 ∆𝑐

Dividing by 𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐 and taking logs yields:

̇ −𝜃
𝑐(𝑡)
[𝑐(𝑡) ]∆𝑡
𝐵𝑒 −𝛽[𝑡+∆𝑡] [𝑐(𝑡)𝑒 ∆𝑐] 𝑒 [𝑟(𝑡)−𝑔−𝑛 ]∆𝑡 ∆𝑐
𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐
=
𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐 𝐵𝑒 −𝛽𝑡 𝑐(𝑡)−𝜃 ∆𝑐
̇ −𝜃
𝑐(𝑡)
[𝑐(𝑡) ]∆𝑡
𝑒 −𝛽∆𝑡 [𝑐(𝑡)𝑒 ∆𝑐] 𝑒 [𝑟(𝑡)−𝑔−𝑛 ]∆𝑡 ̇
𝑐(𝑡)
−𝜃
[𝑐(𝑡) ]∆𝑡
1= = [𝑒 ∆𝑐] 𝑒 [𝑟(𝑡)−𝑔−𝑛 −𝛽]∆𝑡
𝑐(𝑡)−𝜃
̇
𝑐(𝑡)
−𝜃[
𝑐(𝑡) ]∆𝑡
=𝑒 ∆𝑐 −𝜃 𝑒 [𝑟(𝑡)−𝑔−𝑛 −𝛽]∆𝑡
̇
𝑐(𝑡)
0 = −𝜃 [ ] ∆𝑡 − 𝜃 ln ∆𝑐 + [𝑟(𝑡) − 𝑔 − 𝑛 − 𝛽]∆𝑡
𝑐(𝑡)
̇
𝑐(𝑡)
𝜃[ ] = −𝜃 ln ∆𝑐 + 𝑟(𝑡) − 𝑔 − 𝑛 – 𝛽
𝑐(𝑡)
̇
𝑐(𝑡) −𝜃 ln ∆𝑐 + 𝑟(𝑡) − 𝑔 − 𝑛 − 𝛽 𝑟(𝑡) − 𝑔 − 𝑛 − 𝛽
= = − ln ∆𝑐
𝑐(𝑡) 𝜃 𝜃
The choice of 𝑐(0) is then determined by the requirement that the present value of lifetime
consumption over the resulting path equals initial wealth plus the present value of future earnings.

1.2 The Dynamics of the Economy

1.2.1 The Dynamics of 𝒄
Taking account Euler’s equation, the evolution of 𝑐 not just for a single household but for the
economy as a whole:
̇
𝑐(𝑡) 𝑟(𝑡) − 𝜌 − 𝜃𝑔 𝑓′(𝑘(𝑡)) − 𝜌 − 𝜃𝑔 𝑓′(𝑘(𝑡)) − 𝜌
= = = − 𝑔 → 𝑟(𝑡) = 𝑓′(𝑘(𝑡))
𝑐(𝑡) 𝜃 𝜃 𝜃
̇ = 0 → 𝑓 ′ (𝑘(𝑡)) = 𝜌 + 𝜃𝑔 to a capital level of 𝑘(𝑡)∗ = 𝑘(𝑡)
𝑐(𝑡)
̇ <0
𝑘(𝑡) > 𝑘(𝑡)∗ → 𝑓 ′ (𝑘(𝑡)) < 𝜌 + 𝜃𝑔 → 𝑐(𝑡)
̇ >0
𝑘(𝑡) < 𝑘(𝑡)∗ → 𝑓 ′ (𝑘(𝑡)) > 𝜌 + 𝜃𝑔 → 𝑐(𝑡)

1.3 The Dynamics of 𝒌

Actual investment is output minus consumption, 𝑓(𝑘(𝑡)) − 𝑐(𝑡)
̇ = 𝑓(𝑘(𝑡)) − 𝑐(𝑡) − (𝑛 + 𝑔)𝑘(𝑡)
𝑘(𝑡)
̇ is zero when consumption equals the difference between the actual output and break-even
𝑘(𝑡)
investment lines.

The value of 𝑐 is increasing in 𝑘 until 𝑓 ′ (𝑘(𝑡)) = 𝑛 + 𝑔 (the golden-rule level of 𝑘) and is then
decreasing
̇
𝑘(𝑡) 𝑓(𝑘(𝑡)) 𝑐(𝑡)
= − − (𝑛 + 𝑔)
𝑘(𝑡) 𝑘(𝑡) 𝑘(𝑡)
̇ = 0, 𝑘 is falling; when 𝑐 is less than this level, 𝑘 is rising.
When 𝑐 exceeds the level that yields 𝑘(𝑡)
1.4 The Phase Diagram
Recalling that 𝑘 ∗ is defined by 𝑓 ′ (𝑘(𝑡)∗ ) = 𝜌 + 𝜃𝑔, and that the golden-rule 𝑘 is defined by
𝑓 ′ (𝑘(𝑡)𝐺𝑅 ) = 𝑛 + 𝑔. Since 𝑓 ′′ (𝑘(𝑡)) is negative, 𝑘(𝑡)∗ is less than 𝑘(𝑡)𝐺𝑅 if and only if 𝜌 + 𝜃𝑔 is
greater than 𝑛 + 𝑔. This is equivalent to 𝜌 + 𝑛 − (1 − 𝜃)𝑔 > 0 , which we have assumed to hold so
̇ = 0 curve.
that lifetime utility does not diverge. Thus 𝑘(𝑡)∗ is to the left of the peak of the 𝑘(𝑡)