World Income Per Capita: 0 - 2000

GDP Per Capita

6300

GDP Per Capita (1990 Int'l $)

5300

4300

3300

2300

1300

300
0

250

500

750

1000

1250

1500

1750

2000

Regional Income Per Capita: 0 - 2000

GDP Per Capita (1990 Int'l $)

24000

20000

16000

12000

8000

4000

0
0

250

500

Western Europe
Latin America

750

1000

1250

Western Offshoots
Africa

1500

1750

2000

Asia
Eastern Europe

The Great Divergence:

Income Per Capita
28000

Western Offshoots

24000

GDP Per-Capita

20000

Western Europe

16000
12000

8000

Latin America
Eastern Europe
Asia
Africa

4000

0
1820

1870

1913

1950

1998

Inequality in the World Economy: 02000

Year

Western
Offshoots

Western
Europe

Latin
America

Asia

Africa

Ratio:
Rich/Poor

400

450

400

450

425

1.1 : 1

1000

400

400

400

450

416

1.1 : 1

1820

1202

1,204

692

581

418

3: 1

2001

26,943

19,265

6,150

3861

1,489

18 : 1

Exogenous Growth Model

(Based on Solow 1956)

Constant Technology
Discrete time: t 0, 1, 2, . . . ,
Two factors of production:
L t - Labor
K t - Capital
Produce one final good that can be used
for consumption or as capital in the
production process.

Factor supply
Labor supply at t 1 :
L t1 1 nL t
where:
L 0 is given
n 1
capital supply at t 1 :
K t1 S t 1 K t
where:
K 0 is given
S t - aggregate saving
0, 1.

A1:
n 0

Production
output produced at time t :
Y t FK t , L t
A2:
F k K t , L t , F L K t , L t 0,
F KK K t , L t , F LL K t , L t 0, for all K t , L t 0
lim k0 F k K t , L t
lim k F k K t , L t 0
F0, L t 0
FK t , L t FK t , L t

Y t FK t , L t L t FK t /L t , 1 L t fk t
where k t K t /L t

It follows from A2:

f0 0 L t f0 F0, L t 0
for all k t 0 :
f k t F K K t , L t 0
F K K t , L t dL t fK t /L t /dK t f k t
and
f k t L t F KK 0
F KK df K t /L t /dK t f k t /L t
lim k t 0 f k t lim k t f k t 0

Moreover:
since:
FK t , L t FK t , L t , differentiating with
respect to :
FK t , L t F K K t F L L t
and dividing by L t :
fk t f k t k t F L

fk t f k t k t F L 0

Remark:
In a competitive environment:
the rate of return per unit of capital (rental
rate):
F K f k t
the wage rate per unit of labor:
F L fk t f k t k t

Remark:
Since FK t , L t F K K t F L L t , it follows
from differentiating with respect to L t that
F L F KL K t F LL L t F L

F KL K t F LL L t 0

F KL 0

Consumption, Saving and Investment

S t sY t
where s 0, 1

Capital Accumulation:
K t1 S t 1 K t
sL t fk t 1 K t

k t1

sL t fk t 1 K t L t
K
t1

L t1
Lt
L t1

k t1

sfk t 1 k t

k t
1n

The Dynamical System

k t 0 such that
k t1 k t t
where k 0 is given

Let y t be output per worker

y t Y t /L t fk t

k t 0 uniquely determines y t 0

Properties of k t :
0 0

sf
k t 1

k t
0 k t 0
1n
k
sf
t
0 k t 0
k t
1n

lim k t
k t 0

lim k t 1 0, 1
k t
1n

Remark:
The strict concavity of k t follows from:
1. the strict concavity of fk t
2. saving is a constant fraction of output

Steady states
k such that:

k k sfk 1 k
1n

n k sfk
there exist 2 steady states:
k 0 unstable
k 0 stable
Remark:
n sf k

Comparative Statics
Proposition.
dk 0
dn
dk 0
ds
dk 0
dk 0

Proof.
Let
Gk , n, s n k sfk 0

dk
dn
dk
ds

G
n
G
k
G
s
G
k

n sf k
fk

n sf k

initial condition do not matter since there

exists a unique globally stable steady state
equilibrium

Comparative Dynamics
Let
k t k t1 k t
kt
Proposition.
d k t
0
dn
d k t
0
ds
d k t
0
dk t

Proof.
sfk t 1 k t
k t /k t
1n
sfk t n k t

1 nk t
sfk t

n
1n
1 nk t

kt

d k t
sfk t
1 0

dn
1 n 2
1 n 2 k t
d k t
fk t

0
ds
1 nk t
d k t
s

fk

f
k t k t 0
t
2
dk t
1 nk t

Conclusion: no growth in the long-run

without technological progress
Testable Implications and Evidence
conditional convergence
convergence
convergence

35

36

Threshold Externalities
Y t A t FK t , L t A t L t fk t
where k t K t /L t
and A t Ak t
A H if k t k
Ak t
AL

if k t k

Dynamics:
H k t if k t k
k t1
L k t if k t k

Testable Implications
club convergence

Endogenous Growth
Ak Basic Model
(Based on Rebelo JPE 1991)
Production:
Y t AK t y t Ak t
where A n /s
The Dynamical System:
sfk t 1 k t
k t1
1n
sAk t 1 k t

1n
sA 1 k t k t
1n

therefore:
0 0
k t

sA 1
1n

1 k t 0

k t 0 k t 0
growth rate:
k t k t1 k t
kt
sfk t 1 k t

k t /k t
1n
sA n
1n
sA n
1n
1n


d k t
0
dn
d k t
0
ds
d k t
0
dk t

no conditional convergence

Growth Accounting
(Solow 1956)
CRS Production
Y AFK, L

Y A Y K Y L Y
A
K
L
j - the change in the variable between
two periods, j Y, A, K, L.

Y A Y K Y L Y
A
K
L

Y A K Y K L Y L
Y
K K Y
L L Y
A

A Y S K K S L L
K
L
Y
A
S K - the share of capital the elasticity of
output with respect to capital
S L - the share of labor the elasticity of
output with respect to labor

y Afk
y Y/L
k K/L

A y S K k
y
A
k

Overlapping Generations Model

(Based on Diamond 1965)
Production
Y t FK t , L t
F satisfies A2

Y t FK t , L t L t Fk t , 1 L t fk t ,
where k t K t /L t
Wage per worker
w t fk t f k t k t
Return to capital ( 1 intrest rate between
t 1 and t
R t f k t 1

Individuals
A generation of size L is born every period
and lives for two periods
Individuals:
supply labor inelastically, consume and
save in their first life period
consume in the second
Utility of the working generation:
u t uc yt , c ot1
Budget constraint
c ot1 R t1 s yt R t1 w t c yt

c ot1
y
wt
ct
R t1

Optimization:
s yt sw t , R t1
since consumption (in second period) is
normal:
ds t 0
dw t
A3
ds t 0
dR t1

The evolution of capital

K t1 Ls yt

k t1 sw t , R t1
sfk t f k t k t , f k t1 1
under A3,
k t1 k t
Properties of k t :
k t 0
(follows from the normality of consumption)
0 0

k t fk t

Comment:
Aggregate saving per worker in
the economy
s t s yt s ot
where
s ot 1 k t
s yt k t1

s t s yt s ot k t1 1 k t
in the steady state:
s t s yt s ot k

Cobb-Douglas production and utility

Production
Y t FK t , L t AK t L t 1 L t Ak t
k t K t /L t
wage per worker
w t 1 Ak t
return to capital
R t Ak 1
1
t
utility
u t uc t , c t1 ln c t 1 ln c t1
1
optimization:
st 1 wt
2

The evolution of capital

1 1 Ak k t
k t1 s t
t
2
Properties of k t :
0 0
k t 0 k t 0
k t 0 k t 0
lim k t
k t 0

lim k t 0

k t

steady state
k

1 A
2

1/1

Human Capital Accumulation

(Related to Uzawa IER 1965, Barro JPE
1990)
Production:
Y t FK t , H t
satisfies A2.
Ht Ltht
where L t L 1, and h t1 he t e t
note that this implies that human capital
fully depreciate at the end of each period

output per capita.

Y t y t h t fk t
where k t K t /H t K t /h t
Total investment in physical and human
capital is
S t sy t sh t fk t

Assumption: physical capital fully

depreciate at the end of each period
( 1)
Allocation of investment: a fraction of
savings is allocated to physical capital and
a fraction 1 to human capital.
k t1 sAk t h 1
t
h t1 1 sAk t h 1
t


y t1 Ak t1 h 1
t1

1
AsAk t h 1
t 1 sAk t h t

A 1 1 sAk t h 1
t
A 1 1 sy t
for sufficiently high: productivity of
education , saving rate s, and productivity
A there exists an efficient allocation such
that .
A 1 1 s 1
No convergence
No limit to human capital accumulation

Endogenous Technical Change

(Based on Frankel AER 1962, Romer JPE
1986, Lucas JME 1988)
The level of technology is:
A t Ak t
is external to the firm
(in Lucas 1988 Ah is a function of human
capital)
Production:
y t Ak t fk t
where fk t is derived from a function
FK t , L t that satisfies A2

Saving per worker is sy t , s 0, 1.

Suppose Ak t fk t is linear in k t
example:
Ak t k 1
t
fk t Bk t

y t Ak t fk t Bk t
Ak model with constant factor shares.

Endogenous R&D
Quality Ladder Model
(Related to Lucas JME 1988, Grossman
Helpman 1991; Aghion Howitt
Econometrica 1992)

Production of the final good:

The final good produced by each worker in
the final good sector is
yt At
where,
A t A t1 i t
A t1 is the non-excludable existing
technology and i t is new knowledge
(inventions) purchased by the worker.

Individuals
In each period a population of size N joins
the economy
Individuals are active one period in which
they work in the final good sector or in the
R&D sector
The number of workers in the R&D sector
is H t
The number of workers in the final good
sector (production) is L t
Lt Ht N

Production of technology:
The number of non-rival inventions each
worker in the R&D sector produces in t is:
A t1
inventions are made at the beginning of
the period and sold to producers
Equilibrium
In equilibrium all workers purchase all
inventions:
i t A t1 H t

A t A t1 A t1 H t A t1 1 H t
The surplus generated by each invention
used by each worker is 1.

The surplus is divided between production

workers and R&D workers: a fraction
0, 1 is allocated to the R&D worker
and 1 to the production worker.
income of each R&D worker in t is
I Ht A t1 L t
income of each production worker in t is
I Lt A t1 1 A t1 H t

for L 1, equilibrium in the labor market

(individuals are indifferent between the two
occupations) implies:
I Ht I Lt
A t1 L t A t1 1 A t1 H t
L t 1 1 H t

1 1 H t
Lt

1 H t H t
N Ht

1 H t
N

 if N 1, for all t:
H N 1/
L 1 N 1/
if N 1
H0
LN
Since A t A t1 A t1 H
g t g A t A t1 H
A t1
N 1 if N 1

if N 1

Conclusions
1. Growth is affected by:
scale
R&D productivity
patents property rights
2. Crucial elements:
technology is non-rival and excludable
linearity of technological progress with
respect to the technological level

Comments
1. monopolistic competition
may generate over investment in R&D
2. externality to technology
may generate under investment in R&D
3. if investment in technology takes place
before benefits from the technology are
exhausted
the interest rate/time preference have an
effect on R&D investment

Criticism
(Jones 1995)
1. Economies of scale
2. Non decreasing productivity in R&D
inconsistent with empirical evidence from
the 20th century
3. New technology is proportional to the
stock of old technology

Define t
t A t A t1 A t
At
At
Suppose
A t1 A t A t
A t gR&DA t
1 gR&DA t
where R&D is constant over time
(it can be replaced by your favorite
candidate:
human capital, population, or anything
else)


t gR&D
and
A t gR&DA t

Suppose, in contrast

A t gR&DA t 1
1
t A t gR&DA t
1
At
if 1, t is growing over time
converging to infinity
if 1, t is declining over time
converging to zero

Scale Effect in a Malthusian

Economy
(based on Kremer 1993)
Production

A
X
t
Y t A t X L 1
Lt
t
Lt
where L t is the adult population in t, X is
the constant land size, augmented by a
productivity coefficient, A t .
income per adult individual is

A
X
t
y t Y t /L t
Lt

Individuals
live two periods: childhood and adulthood.
adults work, consume and raise children
Preferences:
u t 1 log c t log n t 0, 1
c t - consumption in the household
n t - number of children
Budget constraint

A
tX
c t n t
Lt
is the cost of raising a child

Optimization
nt AtX
Lt
where /

The evolution of population

L t1 n t L t y t L t
Y t
A t X L 1
t
for any given A there exists a unique
globally stable steady state ,
L 1/ AX

Suppose A t evolves sufficiently slow

the economy is at the proximity of the
Malthusian equilibrium:
L t 1/ XA t
Consider population dynamics under:

1. Technological progress is constant

A t1 A t g
At

L t1 1 gL t
where L 0 is given.

L t 1 g t L 0
lnL t lnL 0 t ln1 g
Prediction: log population evolves
linearly over time.

2. Technological progress is increasing

with population size
A t1 A t gL ; g L 0
t
t
At
Prediction: log population is a convex
function of time.

Evidence (from million BC until the 20th

century)
Consistent with #2
Interpretation:
A larger population generates more
non-excludable inventions.
A growing population allows for increasing
scope for division of labor.

Inequality and Growth

The credit market imperfection
approach
(Galor and Zeira 1993)

Individuals
OLG
A generation of size 1 is born every period
and lives for two periods
Each individual has one parent and one
child

in their first life period:

Individuals are endowed with a parental
bequest, invest in human and/or physical
capital,
in their second life period:
Individuals supply labor inelastically,
consume and bequeath

Preferences of individual i born in t are

defined by the utility function:
u it 1 log c it1 log b it1
where 0, 1.
Budget constraint
c t1 b it1 I it1

b it1 bI it1 I it1

The production of human capital

there is an indivisible cost, h, invested in t
(in the first period of life) to become skilled
in t 1

Capital markets and prices

unrestricted international capital flows at
the world gross interest rate R.
the gross interest rate for borrowers for
investment in human capital is
R
where 1.

Wages
w st w s
as follows from the production function
w ut w u
A1:
w s w u hR
A2: is sufficiently large such that:
w s w u hR

Investment decisions and income

if b it h
I it1 w s b it hR
if b it h
I it1 maxw s h b it R, w u b it R

where, as follows from A1 and A2 there

exists
b w u w s hR 0, h
R 1
such that
w s h b R w u b R
and individuals choose to invest in human
capital if and only if b it b .

alternative presentation: the cost of

education, which is strictly decreasing in b it
for b it h, is equal to the return,
h b R b R
hR b 1R
ws wu

The dynamical system

b it1
for b it b
w s h b it R for b it b , h
w u b it R

w s b it hR
b it

for b it h

A3: R is sufficiently small and w s is thereby

sufficiently large such that
R 1
w s h
A4: w u is sufficiently small such that
w u b 1/ R
implying that
w u b R b

Note that:
(1) this assumption can be expressed as
an assumption on the parameters:
w u 1 RhR w s /R 1
(2) A4 implies that
R 1

Assumptions A1 - A4 assure that the

dynamical system is characterized by 2
stable steady states:
u
w
bL
1 R
i.e., b L b L R w u
s hR
w
bH
1 R
i.e., b H b H hR w s
and a threshold unstable steady state
s hR
w
bT
1 R
i.e., b T b T hR w s

Replacing the non-convexities of

the technology
(Moav 2002)
Production
Y t FK t , H t
H t - efficiency units of human capital.
Individuals
as in the Galor-Zeira model, with the
following utility function:
u it 1 log c it1 log b it1 ,
where 0, 1 and 0.

budget constraint:
I it1 b it1 c it1

b it1 bI it1
if

I it1 ;

I it1 if

I it1 ;

where 1 /.

Capital markets
unrestricted international capital flows at
the world capital rate of return, R, uniquely
determine the wage per efficiency unit of
human capital w.
(alternatively: Y t wH t RK t
individuals can not borrow for sake of
investment in human capital.

The formation of human capital

the level of human capital is an increasing
concave function of real resources
invested in education
1 e it if e it e;
h it1 he it
1 e
Assumption:
w R

if e it e.

The evolution of income

I it1

w1 b it

if

b it e;

w1 e Rb it e

if

b it e.

since
if

I it1 ;

if

I it1 ;

if

I it 0;

w1 I it

if

I it 0, e;

if

I it e,

b it1

0
I it1

I it1

w1 e
RI it e

I it

Assumptions:
1.
w
2.
w1 e e

e
3.
R 1

w
w 1

 there exists an income threshold, ,

1w

w 1
the dynamical system, I it1 I it ,
generates multiple steady states. A
poverty trap, I L w, a high income steady
state, I H , and a threshold income
I L , I H , where,
we 1 R e
H
I
.
1 R

Robustness and endogenous

wages
(based on Banerjee and Newman 1993)
Basic Galor Zeira model with random
shock to income and endogenous wages
Random shocks
a fraction 1 of individuals who did not
invest in human capital become skilled
workers
a fraction 2 of individuals who did invest in
human capital are unskilled workers

 1 and 2 are small and, for sake of

simplicity, are ignored by individuals when
making investment decisions.
Assumption A3 assures that the positive
shock places individuals in the basin of
attraction of the high income steady state
and it is assumed that the negative shock
places individuals in the basin of attraction
of the low income steady state. (this
assumption holds for sufficient low values
of , R and w u .

The fraction of skilled individuals in t 1

H t1 H t 1 2 1 1 H t
1 1 1 2 H t .
the steady state value of H,
H 1
1

is independent of the initial wealth

distribution

Endogenous wages
output in the agricultural sector, Y At ,
Y At F A T, L t
where, T is the constant land size, F A is a
CRS production function characterized by
decreasing positive marginal products and
boundary conditions that assure an interior
solution to the producers maximization
problem.

w u w u L t
dw u L t /dL t 0, and lim L t 0 w u L t

it is assumed that
w u 1
where is the threshold level of w u above
which the dynamical system describing the
evolution of b it is characterized by one
steady state
there exists L such that w u L .

 for w u , is charctarized by a poverty

trap
for w u , is charctarized by a unique
steady state
Assume that
2
L

L
1 2
where L Therefore:
If L 0 L the economy will converge to the
steady state:
L L , H 1
1

If L 0 L (and 2 is sufficiently small) the

dynamical system governing the evolution
of b it is characterized by a unique high
income steady state. In equilibrium
therefore a sufficient fraction of the
population will have a bequest b h and
the net return to education declines to
zero, uniquely determining the number of
uneducated workers in the steady state, L :
w s w u L hR.

Replacing in the dynamical system:

b it1 w u L b t R w s b t hR
which is linear in equilibrium and has a
unique high income globally stable steady
state. In the steady state all individuals
converge to the high income steady state,
but they are subject to negative income
shocks that place them in a lower point
along the dynamical path. In the steady
state H 1 L .

Income mobility
(based on Maoz-Moav, 1999)
Production and wages
Y t FL t , E t
E t - the number of skilled (educated)
workers
L t - the number of unskilled workers
Et Lt 1
w et w e E t and w ut w u E t

Individuals
OLG
A generation of size 1 is born every period
and lives for two periods
Each individual has one parent and one
child
First life period:
receive b t from parent, consume and
invest in human capital
Second life period:
supply labor inelastically, consume and
bequeath

Preferences
u it log c it 1 log c it1 log b it1
c it - consumption when young
c it1 - consumption when old
b it1 - transfer to the child

The production of human capital

there is an indivisible cost, h it , invested in t

to become skilled in t 1

Capital markets

do not exist

Optimization
The budget constraints of individual i are:
c it i h it b it
c it1 b it1 w it1
where,
i

1 i acquires education
0

otherwise

w it1

w et1

i 1;

w ut1 i 0.

Optimization in second period:

zw it1 max1 log c it1 log b it1
s. t. c it1 b it1 w it1

b it1 w it1

zw it1 log w it1 1 log1 log

Optimization in first period:

individual i will choose to invest in human
capital if and only if:
logb it h it zw et1 log b it zw ut1

logb it h it logw et1 log b it logw ut1

b it h it w et1 b it w ut1

b it

w et1 w ut1
w et1

i
h t h it

i
h t - the critical value of the cost of
i
i
education such that i will invest iff: h t h t
i i
i
i
e
h t /b t 0, h t /w t1 0, h t /w ut1 0.

e
w ut1
e
h t w t 1 e
w t1
u
w ut1
u
h t w t 1 e
w t1

Fh - the distribution of h i which is

independent of parental education and
time.

e
u
E t1 E t F h t 1 E t F h t
Since the left hand side is strictly
increasing in E t1 and the right hand side is
strictly decreasing in E t1
E t1 is uniquely determined by E t :
E t1 E t
where E t is increasing in , and w e ,
whereas w u has an ambiguous effect.

Example: in the handout

The Evolution of Total Fertility Rate

Across Regions, 1960-1999
7

1
1960

1964

Western Offshoots

1968

1972

Europe

1976

1980

Africa

1984

Asia

1988

1992

1996

Latin America

Oceiania

The Demographic Transition in Western

Europe: Total Fertility Rates
6

T o ta l F e r tility R a te

1
1851-1855

France

1876-1880

1901-1905

1926-1930

Netherlands

England and Wales

Germany

1951-1955

Norway

1976-1980

Sweden

Finland

view, across countries, and across individuals. An early discussion of this fact
appears in the seminal article on fertility choice by Becker (1960). Indeed this
puzzling correlation was one of the main impetuses to Beckers early work.10
Quoting from Becker (1960) (p. 217): Indeed, most data tend to show a negative
relationship between income and fertility. This is true of the Census data for 1910,
1940 and 1950, where income is represented by fathers occupation, mothers
education or monthly rental; the data from the Indianapolis survey, the data for
nineteenth century Providence families, and several other studies as well.11
Figure
1: Fertility
Incomeinin
2000
Dollars
Figure
3: CEBby
vs.Occupational
Occupational Income
2000
Dollars
6.5
Birth Cohort

6.0

1828
1848
1868
1888
1908
1928
1948

Children ever Born

5.5
5.0
4.5
4.0

1838
1858
1878
1898
1918
1938
1958

3.5
3.0
2.5
2.0
1.5
0

20,000

40,000

60,000

80,000

Occupational Income in 2000 Dollars

Source: Jones and Tertilt (2008)

In a recent study, Jones and Tertilt (2008) use U. S. Census Data on lifetime
fertility and occupations to document this negative cross-sectional relationship
in the United States.12 They find a robust negative cross-sectional relationship be10

See the discussion in Hotz, Klerman, and Willis (1993).

The studies Becker is referring to are U.S. Census (1945), U.S. Census (1955), Whelpton and
Kiser (1951), and Jaffe (1940).
12
Income is based on the median annual income for a given occupation in 1950 and adjusted for
11

Endogenous fertility
(Moav, 2005)
Production
Y t wH t
where H t is the aggregate level of human
capital in t.
Individuals
live two periods:
In childhood they acquire human capital
in adulthood they work consume and raise
children

preferences:
u it 1 log c it log n it h it1
c it - consumption of the household
n it - the number of children
h it1 - the level of human capital of each
child, measured in efficiency units

The formation of human capital

h it1 he it ,
e it , is measured in efficiency units of labor
the real cost of the investment in
education is we it

Clarifying note: suppose it is the time a parent spends

educating each of his children. Therefore the effective
level of education is e it it h it , and the cost is
therefore it h it w, hence the cost of e it is we it

h0 1, h 0 , h e it 0, h e it 0.

Budget constraint
n it wh it e it c it wh it whe it1
is the minimum time cost required for
raising a child
A1:

1/

Optimization
Consumption:
c it 1 wh it 1 whe it1
Education:
if:
h0
1 h 0

h it
h it

e it 0
otherwise e it 0, is given by
he it
he it
e i

h
t
i
i
i
i
h t e t
he t1 e t

noting that
he it
d
h it e it

/de it 0

he it
e i

h
t
h it e it

The marginal return to quality and quantity.

h(et )
h(et 1 ) + et

1
h(et 1 )

h' (et )

(et 1 )

et

there exists a single valued function e it

such that,
0 if e it
e it1 e it
0 if e it
where e it 0 for e it and 0 is
unique and given by 1/ h.

et +1

450

(et )

e eT

et

There exists a human capital function that

generates multiple steady states.
Consider for instance the following human
capital production function,
1 e it if e it ;
h it1 he it
1
where 1 / 2 .

if e it .

Hence,
if

e it ;

0, if

e it ;

e it ,

0
e it1 e it

if

where 1 / 2 0 and therefore

. The dynamical system e it is
characterized by two stable steady states,
e and 0, and an unstable steady state ,
which is the threshold level of education.

Fertility in the steady states:

n /

for e 0

n 1 /1 for e
where
/ 1 /1

Two Dimensional Dynamical

System
An Example

y t Ak t h t ;

0, 0, 1
k t1 sy t

h t1 h t y t
where:
0,

s 0,

s 1 1

 the dynamical system:

k t1 sAk t h t k t , h t

h t1 h t Ak t h t

Ak t h t

k t , h t

where
k k t , h t 0,

h k t , h t 0

kk k t , h t 0,

hh k t , h t 0

The kk Locus
Let kk be the locus of all pairs k t , h t such
that k t is in a steady-state:
kk k t , h t : k t1 k t :

k sAk h t

/1

k sA 1/1 h t

k h t

if k t k h t , then k t1 k t , h t k t
The kk Locus consists of all the pairs
k h t , h t
1 /1 1
k h t is increasing and concave

The hh Locus
Let hh be the locus of all pairs k t , h t such
that h t is in a steady-state:
hh k t , h t : h t1 h t :
h Ak t h
h

1/1 /1
kt
A

h k t

h t h k t , then h t1 k t , h t h t
The hh Locus consists of all the pairs
h k t , k t
If 1 /1 1
h k t is increasing and concave

As follows from the properties of the

dynamical system, the non-trivial steady
state levels of k h t and h k t are unique
and globally stable.

The Political Economy Approach

Alesina-Rodrik 1994, Persson-Tabellini
1994, Benabou 2000
(the model is based on Benabou 2000)

One period endowment economy

Wealth distribution
Cumulative distribution of (pre-tax) wealth:
Fx

for

x0

/1 x

for

x 0, 1 /

for

x 1 /

where 0, 1
The density function:
0
fx

1
x

for

x0

for

x 0, 1 /

for

x 1 /

1.2

0.8

a=0.3
a=0.7
a=0.9

0.6

0.4

0.2

0
0

0.5

1.5

2.5

3.5

1.8

1.6

1.4

1.2

1
a=0.3
a=0.7
a=0.9

0.8

0.6

0.4

0.2

0
0
-0.2

0.5

1.5

2.5

3.5

Mean income:
x

1/

1/

xfxdx

x 1

x dx

1/
0

Median income, m
Fm 1/2
Hence,
Fm

m 1
2

and therefore
m.
m 1 1/
2

Properties of m :
lim m 0
dm
d

m1

1 log 2
2 1/ 3

if
0 if
if

0
0, 1
1

0
0

Indices of equality:
1. Median/Mean ratio:
median m m
mean
x
the higher is the median, that is the
higher is , the lower is inequality.

2. Income variance:
2

varx Ex

x 2
1/

1/

x 2 fxdx 1

x 1 dx

1/

1
2
x

1
2
0
1

2
the variance of x is decreasing in , that
is the higher is , the lower is inequality
Note that for 1, varx 1/3, equal to
the variance of a uniform distribution with a
range of 2 (the variance of the uniform
distribution is given by r 2 /12, where r is the
range), and for 0 varx
Equality increases with
is a measure of equality.

Redistribution
Post-tax Income of individual i, x i :
x i 1 x i x
- fraction of wealth taxed and
redistributed equally among individuals
1 - distortionary taxation
1 - beneficial taxation
Redistribution is preferred by i if
x i x i x x i
for 1
x i x i x x i

Since x 1, redistribution is beneficial for i

if
x i x i x i
redistribution is supported by a fraction
F of the population:
F

for

/1

for

0, 1 /

for

1 /

For 0, 1
F

if 0

F/ 0 if 0, 1
F

/2

if 1

Note that:
For 0 :
lim 0 m 0 and thus F 1 since
0.
For 1 :
m x and thus F 1/2 since 1.
Hence for distortionary taxations:
More equality reduces the pressure for
redistribution.
For 1, 2
F 1/2

F/ 0 for
F/ 0 for
F2 1

for

low
high
expln 4 1/2, 2
1

1
0.9

=2

0.8
0.7

=1.5

0.6
0.5

=1

0.4
0.3
=1/2

0.2
0.1
0
0

0.2

0.4

0.6

0.8

The non-monotonic impact of inequality

captures two effects:
1. More inequality increases the
proportion of less than average income
individuals who support redistribution.
2. More inequality increases the cost of
redistribution for high income
individuals who object redistribution
Inequality and Growth
The theory predicts that inequality:
1. Has a negative effect on growth if
taxation is distorting 1.
2. Has a negative effect on growth if
1. 213, and inequality is sufficiently
low (high .

A two period model of

Geography, Transparency and Institutions
(Mayshar-Moav-Neeman)

We study the impact of transparency on

two standard elements in the (implicit)
principal-agent contract:
Stick threat of dismissal
Carrot share of output

and explain regional differences in

Institutions (private vs. state owned land)
State capacity
State concentration (center vs. periphery)

The principal-agent model

The principal designs the contract to
maximize its expected income
Agents are risk neutral and choose their
effort level to maximize their expected
welfare
The economy operates during two periods:
1 and 2

Output (per agent, in each period)

Y

H if e h and G
L

otherwise

e h, l - effort
G, B - state of nature
p 0, 1 - the probability that G

Information
, B - a public signal about the state
G
of nature
Signal accuracy q 1/2
|G PrB |B
q PrG
|B PrB |G
1 q PrG
is observed after effort decision

Interpretation of the signal

a. Observation of output in other plots
provides information about the state of
nature at a specific plot depending on the
correlation across plots.

Interpretation of the signal

b. An observable signal, such as the
Nilometer that measures the amount of
water in the Nile.

The cost of maintaining the agent

0 if effort is low e l
if effort is high e h

Assumptions:
L
(low output is larger than the maintenance
cost)
pH L
(effort is efficient)

Agents Income and Utility

I - agents expected income
U I - agents periodic utility when
exerting effort
1 - the agents discount factor
V - the value of the agents employment in
the second period
zero - agents value of unemployment

Incentive scheme - the carrot

The principal pays the agent:
a bonus b 0 if output is high Y H
a basic wage regardless of output

Incentive scheme - the stick

The contract could include dismissal of the
agent at the end of period 1 if:

Y L and G
(otherwise the agent is retained)
x - the cost of replacing the agent
1 p p2
p
x

1p
1p
dismissing the agent when B is dominated by
never dismissing

 Two types of contracts are possible in

period 1
Pure Carrot
(denoted by subscript c)
Stick and Carrot
(denoted by subscript s)
In period 2, only Pure Carrot is relevant

The IC constraint under Pure Carrot

pb c
Minimizing the cost of incentivizing the
agent
b c /p

The value of employment in period 2

V pb c
The cost of employment for the principal
under Pure Carrot
c pb c 2

The IC constraint under Stick and Carrot

pb s p 1 pqV
p1 q 1 pqV

noting that V

bs

1 pq
p

The cost of employment for the principal

under Stick and Carrot
s pb s 1 p1 qx
2 pq 1 p1 qx

There exists a threshold q such that:

s c for q q
q

1 px
1 px p

For q q Pure Carrot

For q q Stick and Carrot
x

p
1p

q 1/2

Intuition: a principal relying on a stick to

incentivize the agent has to incur the cost
of dismissal x with probability
1 p1 q
The expected cost of using the stick
1 p1 qx
is decreasing with the quality of
information q

Expected Income - Pure Carrot

The expected income of the agent
I c pb c 2
The expected income of the principal
c pH L L 2
Efficient outcome
I c c pH L L

Expected Income - Stick & Carrot

The expected income of the Agent
I s pb s 2 pq 2 pq
is decreasing with q

The intuition for the decline of I with q

above q
Holding constant the bonus, b, a higher q
implies a lower probability of dismissal,
increasing the value of employment.
Therefore, as q increases b has to decline
to hold the incentive constraint binding.

Expected Income - Stick & Carrot

The expected income of the principal
s pH L L 2 pq
1 p1 qx
is increasing with q
Lower payment to the agent and lower
probability of paying x

Expected Income - Stick & Carrot

inefficient outcome
I s s pH L L 1 p1 qx
inefficiency declines with q

An Illustrative Calibration
EY pH 1 pL 1 (representing
about 1.5 tons of net grain)
p 0. 75, (a bad harvest occurs about
every 4 years)
x 1, 0. 2
q 0. 625

TotalincomeandAgent'sincome
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.5