thinking about the rise and decline of nations

8/2/2019 Thinking About the Rise and Decline of Nations

1/60

Rise and Decline of Nations 1

Thinking about the Rise and Decline of Nations

Daron Acemoglu

June 2005


2/60


Introduction

Many scholars have noted a pattern of rise and decline;

previously advanced civilizations collapse, while others prosper.

Prominent examples include:

The declines of the Roman Empire, the Egyptian empire,

the Chinese Empire in the ancient times.

The decline of Venice and Genoa and the rise of Englandand the Dutch Republic (and then the decline of the Dutch

Republic).

The decline of Inca and Aztec empires and the rise of the

civilizations in North America in the New World.

Existing theories:

Building up of social rigidities (Olson).

Military over-expansion (Kennedy).


3/60


This Talk

A different perspective on the rise and decline.

Main idea: institutions appropriate (or only marginally costly)

under certain circumstances become more costly later.

Thus a theory of interactions between institutions and

economic opportunities.

Indirectly about:

the costs and benefits of the different economic (political)

systems?

the potential for change and flexibility within given

institutional environments.


4/60


Plan of Talk Two examples of rise and decline.

Caribbean versus Northeast America (within a broader

reversal of fortune among former European colonies).

Venice and Spain versus England and the Dutch Republic. A Model of oligarchy versus democracy.

Key trade-off between protecting the property rights of

incumbents versus creating a level playing field for

non-incumbents.

Dynamic distortion of oligarchy.


5/60


The Reversal of Fortune

Use of urbanization and density of population before European

colonization as a proxy for income per capita and how

advanced pre-colonial civilizations are.

Robust negative relationship between income today and

urbanization in 1500 among the former colonies.

Robust negative relationship between income today and log

population density in 1500.

Not due to any geographic variable, or identity of colonial

power.

When urbanization and population density both included,

population density is the main determinant.


6/60

LogGDP

percapita,

PP

P,

1995

7

8

9

10

ARG

AUS

BGD

BLZ

BOL

BRA

CAN

CHL

COL CRI

DOM DZAECU

EGY

GTM

GUY

HKG

HND

HTI

IDN

IND

JAM

LAO

LKA

MEXMYS

NIC

NZL

PAK

PAN

PER

PHL

PRY

SGP

SLV

TUN

URY

USA

VEN

VNM


7/60

LogGDP

percapita,

PPP,

1995

6

7

8

9

10

AGO

ARG

AUS

BEN

BFA B

BHS

BLZ

BOL

BRA

BRB

BWA

CAF

CAN

CHL

CIVCMRCOG

COL

COM

CPV

CRI

DMADOMDZA

ECU

ERI

ETH

GAB

GHAGIN

GMB

GRDGTM

GUY

HKG

HND

HTI

IDN

I

JAM

KEN

KNA

LAO

LCA

LKA

LSO

MAR

MDG

MEX

MLIMOZ

MRT

MWI

MYS

NAM

NERNGA

NIC

NPL

NZL

P

PAN

PER

PHLPRY

R

SDN

SEN

SGP

SLE

SLVSUR

SWZ

TCD

TGO

TTO

TUN

TZA

UGA

URY

USA

VCT

VEN

VNM

ZAF

ZARZMB

ZWE


8/60


The Timing and Nature of the Reversal

When did the reversal take place?

Not when the Europeans plundered the previously rich

societies or killed of their populations.

In the 19th century, and intimately related to industrialization.


9/60

0

10

20

30

40

50

60

70

80

90

100

Urbanization(%)


10/60

50

100

150

200

250

300

350

400

n

dustrialProduction

percapita


11/60


Plantation Versus Industrialization

Specific example of the nature ande timing of the reversal:

The Caribbean Plantation societies in the 17th and 18th

centuries initially prosperous, but then falling behind

Northeastern United States.

Caribbean plantation societies rich from sugar. Highly

oligarchic societies, dominated by the richest plantationowners. Supported by repressive political institutions.

Relatively efficient for production and processing of sugar for

the plantation owners. But no room for change.

In contrast, the more democratic Northeast United States,

more flexible to take advantage of new economic opportunities.

In fact, 19th century growth in the U.S., fueled by industry and

entrepreneurs not previously part of the ruling elite.


12/60


Institutions and Industrialization

Whether a society has institutions of private property or

extractive institutions may matter much more when new

technologies require broad-based economic participation.

Industrialization is such a process, requiring investments from

a large number of agents who were not previously part of the

ruling elite.

Therefore, there are natural reasons to expect that institutional

differences will matter much more during the age of industry.

We find that there is a strong interaction between institutions

and industrialization: institutions start mattering much more

in the age of industry.

This explains why the reversal took place during the 19th

century.


13/60


14/60


Rise of Europe

Western Europe grows faster than Eastern Europe and Asia

after 1500 using either urbanization rates as proxy for income

or Maddisons estimates for GDP.

When we break things out into Atlantic vs. non-Atlantic

Western Europe, almost all of the faster growth is driven by

growth in Atlantic nations (Belgium, Britain, Denmark,

France, Ireland, the Netherlands, Portugal and Spain).

Same pattern when we look at city growth.

The timing of takeoff of various Atlantic ports consistent with

timing of involvement in Atlantic trade by individual countries.

Among Atlantic countries, early growth in Spain and Portugal,

but short lived. Major growth in Britain and Dutch Rep.

Related to institutional change in Britain and Dutch Rep.


15/60

Urbanization rates

weighted by population

0

5

10

15

20

25

1300 1400 1500 1600 1700 1750 1800


16/60

Urbanization rates


0

5

10

15

20

25

30

1300 1400 1500 1600 1700 1750 1800

A l i d W E A l i d E


17/60

GDP per capita from 1500


0

500

1000

1500

2000

2500

1500 1600 1700 1820


18/60

GDP per capita from 1500


0

500

1000

1500

2000

2500

1500 1600 1700 1820 1


19/60

Average of log population in Iberian Atlantic ports and other cities

(balanced panel)

0

0. 5

1

1.5

2

2. 5

3

3. 5

4

4. 5

5

13 0 0 14 0 0 15 0 0 16 0 0 17 0 0 17 5 0

Iberian cities, not Atlantic ports


20/60

Average of log population in British Atlantic ports and other cities

(balanced panel)

0

1

2

3

4

5

6

1300 1400 1500 1600 1700 1750 1800

British cities not Atlantic ports


21/60


22/60


Interpretation

Caribbean plantation economies did well initially because they

provided the right incentives to the major asset holders insociety.

Similar to Venice in the 15th century or even to Spain after the

discovery of the New World.

But long-run growth requires a process of creativedestruction with new entrance and new blood coming in.

Difficult when the regime dominated by incumbents.

Thus industrialization more likely in Northeast United States.

Also, institutional change supporting sustained economic

growth more likely in places where incumbents weaker;

institutional change in England and the Dutch Republic but

not in Spain.


23/60


Ingredients of Model Construct a simple theoretical model to emphasize and clarify

the trade-offs.

Consider an economy where agents enter entrepreneurship or

production work.

Heterogeneity in entrepreneurship.

Entrepreneurial talent imperfectly correlated over time.

Two types of policy distortion:

Redistributive taxation with incentive costs.

Entry barriers protecting incumbents.


24/60


Trade-off Between Oligarchy and Democracy

Entry in democracy, sclerosis in oligarchy.

Lower investment in democracy.

Worse allocation of talent in oligarchy.

Democracy more equal, oligarchy more unequal (lower wages

higher profits).

Oligarchy gets worse over time as the comparative advantage of

incumbents gets worse.

Oligarchy and democracy creating different types of distortions.

But long-run growth may be slower in oligarchy because of

dynamic costs of entry barriers.


25/60


Understanding Rise and Decline of Nations

oligarchy less harmful initially, even encouraging investment

because less redistribution away from major producers.

but harmful as comparative advantage of oligarchs disappears.

oligarchy particularly harmful when new technologies shift

investment opportunities from insiders to newcomers.

oligarchy less flexible than democracy


26/60


Regime Dynamics When does oligarchy transition to democracy?

Two possibilities:

1. Smooth transition to democracy because of within-elite

conflict (small producers disbanding oligarchy).

2. Conflict over regimes.

Incumbents richer, can maintain the system that serve

their interests.

Path dependence in equilibrium regimes possible.

Caribbean vs. United States in the 18th and 19thcenturies.


27/60


Model

Infinite horizon economy, with the unique non-storable good, y.

Preferences

Uj0 = E0t=0

tcjt , (1)

Assume each agent dies with a small probability , consider the

limit of this economy with 0.


28/60


Model (continued)

Choice between entrepreneurship and production work.

Entrepreneurial talent ajt {AL, AH} with AL < AH.

Either already own an active firm, or set it up (costly when

there are entry barriers).

Each agent starts period t with entrepreneurial talent

ajt {AH, AL}, and sjt {0, 1} which denotes the individual

possesses an active firm.

Agent with sjt = 1 member of the elite.

Each agent takes the following decisions: cjt , ejt {0, 1}.

If ejt = 1, then he also makes investment, employment, and

hiding decisions, kjt , ljt and h

jt , where h

jt denotes whether he

decides to hide his output in order to avoid taxation.


29/60


Model (continued)

Three policy choices: a tax rate t on firms, lump-sum transfer,

Tt, and a cost Bt to set up a new firm (pure waste).

Production function for talent ajt :

11

(ajt)(kjt )

1(ljt ),

To simplify assume that ljt = , and that entrepreneur himself

can work in his firm as one of the workers.

Denote: bt Bt/.


30/60


Model (continued)

Denote the wage rate by wt 0.

Profit function (without hiding):

t, kjt , a

jt , wt

=

1 t1

(ajt)(kjt )

1(ljt ) wtl

jt k

jt , (2)

With hiding:

t, kjt , ajt , wt

= 1 1

(ajt )(kjt )1(ljt ) wtljt kjt.

Thus

t ,

Labor market clearing: 10 ejtdj = jSEt dj 1, where SEt isthe set of entrepreneurs at time t.

For agents with sjt = 0, setting up a new firm may entail an

additional cost Bt because of entry barriers.


31/60


Model (continued)

Law of motion of individual states:

sjt+1 = ijt , (3)

ajt+1 =

AH with probability H if ajt = A

H

AH with probability L if ajt = A

L

AL with probability 1 H if ajt = AH

AL with probability 1 L if ajt = A

L

, (4)

Stationary distribution fraction of high-productivity agents:

M L

1 H + L

.

Assume

M > 1,


32/60


Model (continued)

Timing of events:

1. Entrepreneurial talents,

ajt

, are realized.

2. The entry barrier for new entrepreneurs Bt is set.

3. Agents make occupational choices, ejt.

4. Entrepreneurs make investment decisions kjt .

5. The labor market clearing wage rate, wt, is determined.

6. The tax rate on entrepreneurs, t, is set.

7. Entrepreneurs make hiding decisions, hjt.

where

ajt

shorthand for the mapping at : [0, 1]

AL, AH

,

etc.


33/60


Analysis

Economic equilibrium: subgame perfect equilibrium given apolicy sequence {bt, t}t=0,1,....

Equilibrium investments:

kjt = (1 t)1/ajt. (5)

t, wt | sjt , a

jt

=

1 (1 t)

1/ajt wt. (6)

Tax revenues:

Tt = t (1 t)1

1

jSE

t

ajt , (7)


34/60


Analysis (continued)

Who will become an entrepreneur?

1. Entry equilibrium where all entrepreneurs have ajt = AH.

2. Sclerotic equilibrium where agents with sjt = 1 become

entrepreneurs irrespective of their productivity.

An entry equilibrium will emerge only if the net gain to a

high-skill non-entrepreneur of incurring the entry cost and

setting up a firm (at a given wage) is positive.

This net gain takes into account the future benefit of becoming

an elite protected from competition (as a function of future

entry barriers etc.). Determined by simple dynamic programming taking

equilibrium policies as given.


35/60



Let the value function of a worker of type z as a function of the

sequence of future policies and equilibrium wages, (pt,wt):

Wz

pt,wt

= wt + Tt + C Wz

pt+1,wt+1

, (8)

where the continuation values from time t + 1 onwards are:

CWz pt+1,wt+1 =z max

WH

pt+1,wt+1

, VH

pt+1,wt+1

bt+1

(9)

+ (1 z)max

WL

pt+1,wt+1

, VL

pt+1,wt+1

bt+1

.

These incorporate optimal occupational choice from time t+1

onwards.

Similarly, for a current entrepreneur

Vz

pt,wt

= wt + Tt + z (t, wt) + C V

z

pt+1,wt+1

, (10)


36/60

wt LS

LD

wtL

wtH

wtH+bt

wtL-bt


37/60

wt LS

LD

wtH+bt

wtL

wtL-bt

wtH


38/60



Define the net value of entrepreneurship as a function of an

individuals skill a and ownership status, s,

N V

pt, wt | Az, s

= Vz

pt,wt

Wz

pt,wt

(1 s) bt,

where the last term is the entry cost for agents with s =0.

N V pt, wt | AH, sjt

= 1 N V pt, wt | ajt

, sand

N V

pt, wt | ajt , s

NV

pt, wt | AL, sjt = 0

.

Therefore, high-skill incumbents remain entrepreneurs and

low-productivity workers never become entrepreneurs.

Whether low-productivity incumbents remain entrepreneurs

depends on taxes, wages and entry barriers.


39/60



Define entry equilibrium wage such that

N V

pt,

wHt , wt+1

| ajt = AH, sjt = 0

= 0.

So

wHt

1 (1 t)

1/AH bt

+

CVHpt+1,wt+1 CWHpt+1,wt+1 ,

Similarly, sclerotic wage is

wLt

1 (1 t)

1/AL

+

CVL pt+1,wt+1 CWL pt+1,wt+1 .

An entry equilibrium only when

wHt wLt . (11)

d l f N


40/60



Therefore, in equilibrium wet = wHt .

Define fraction of high-productivity entrepreneurs:

t = Pr

ajt = AH | ejt = 1

= Pr

ajt = A

H |j SEt

Since no entry barriers initially, 0 = 1.

Law of motion of t:

t = Ht1 + L(1 t1) if (11) does not hold1 if (11) holds . (12)

Ri d D li f N i 27


41/60


Political Equilibrium

Consider two simple extreme political regimes:

1. Democracy: the policies bt and t are determined by

majoritarian voting, with each agent having one vote.

2. Oligarchy (elite control): the policies bt and t are

determined by majoritarian voting among the elite at time t.

Focus on Markov perfect equilibria.

Ri d D li f N ti 28


42/60


Democracy

Non-elites in the majority.

Majoritarian voting: taxes will be chosen to maximize per

capita transfers,

Tt (bt, t) =

t(1t)1

1

jSE

t

ajt if t

0 if t >

, (13)

where t is the tax rate expected by the entrepreneurs and t is

the actual tax rate set by voters.

Since 0 profits, entry barriers will be chosen to maximize

equilibrium wages, thus bt = 0. Intuitively, entry barriers reduce labor demand and depress

wages.

Ri d D li f N ti 29


43/60


Democracy (continued)

Proposition 1 A democratic equilibrium always features t =

and bt = 0, and ejt = 1 if and only if a

jt = A

H, and t = 1. The

equilibrium wage rate is given by

wDt =

1 (1 )1/AH,

and the aggregate output is

YDt = YD

1

1 (1 )

1

AH. (14)

Perfect equality.

Ri d D li f N ti 30


44/60


Oligarchy

Policies determined by majoritarian voting among the elite.

To simplify this talk, assume

12

AH

AL+ 1

2, (15)

which ensures that low and high-skill elites prefer low taxes.

Otherwise, low-skill elites side with the workers to tax the

high-skilled elites.



45/60


Oligarchy (continued)

Then entry barriers will be set

bt bEt

AH

1 +

CVH

pt+1,wt+1

CWH

pt+1,wt+1

.

(16)

so as to prevent entry.

Imposing wet+n = 0 for all n 0,

VL =1

1

1

1 H

AL + LAH

(1 (H L))

, (17)

and

VH =1

1

1

1

1 L

AH +

1 H

AL

(1 (H L))

.

(18)



46/60



Using these equilibrium relationships,

bt = bE

1

1

1

1

1 L

AH +

1 H

AL

(1 (H L))

.

(19)

Wages are zero and aggregate output is

YEt = t1

1 AH + (1 t)

1

1 AL (20)

where

t = Ht1 + L(1 t1)

with

limt

YEt = YE

1

1

AL + M(AH AL)

. (21)



47/60



Proposition 2 Suppose that condition (15) holds. Then an

oligarchic equilibrium features t = 0 and bt = bE as given by (19),

and the equilibrium is sclerotic, with equilibrium wages w

e

t = 0,and fraction of high-skill entrepreneurs

t = Ht1 +

L(1 t1) starting with 0 = 1. Aggregate

output is given by (20) and decreases over time starting at

YE0 =1

1AH with limt Y

Et = Y

E

as given by (21).



48/60


Comparison of Oligarchy and Democracy

We always have that initially:

1

1 (1 )

1

AH < YE0 =1

1 AH.

Will oligarchy fall behind democracy?

If (1) democratic taxes are low and not very distortionary;(2) selection of entrepreneurs is difficult, and (3)

comparative advantage in entrepreneurship is important,

then oligarchy ultimately worse than democracy:

Condition for this:

(1 )1

> AL

AH+ M

1 A

L

AH

. (22)


49/60

Output

Output

Output

YD

YE

YE0

YE

Yt



50/60


Comparison of Oligarchy and Democracy(continued)

Workers always worse off in oligarchy than in democracy.

What about entrepreneurs?

High-skill entrepreneurs always better off. But

Proposition 3 If

1 H

AL/AH + L

(1 (H L)) 0 a new technology arrives.

Productivity with new technology:

1

1 (ajt)

(kjt )1(ljt )

,

where > 1

Law of motion of ajt orthogonal to ajt , and given by

a

j

t+1 =

AH with probability H if ajt = A

H

AH with probability L if ajt = A

L

AL with probability 1 H if ajt = AH

AL with probability 1 L if ajt = A

L

, (24)



52/60


New Technologies (continued)

In democracy, aggregate output jumps from YD to

YD 1

1 (1 )

1

AH.

In oligarchy, elites will stay in entrepreneurship despite their

worse comparative advantage

For example, if AL > AH, then aggregate output jumps to

and remains at

YE 1

1 AL + M(AH AL)

,

Potential explanation for why oligarchic societies dont adjust

well to new opportunities/technologies.



53/60


Regime Dynamics

Two cases to consider:

1. Conflict within the elitewhen low-skill elites worse off in

oligarchy than in democracy (when condition (23) holds),

they disband oligarchy when they become the majoritywithin the elite.

2. Conflict between classes over regimesthe elite prefer

oligarchy and the citizens democracy; income distribution

matters for regime dynamics; possibility of path dependence.



54/60

Regime Dynamics (continued)

Suppose that, in oligarchy, current elite can legislate apermanent transition to democracy. Then

Proposition 4 Suppose (15) holds and the society starts

oligarchic.

If (23) does not hold, then for all t the society remains

oligarchic.

If (23) holds, then the society remains oligarchic until date

t = t where t = min t N such that t 1/2 (whereby

t = Ht1 +

L(1 t1) for t < t starting with 0 = 1).

At t, the society transitions to democracy.

The low-skill elites disband the oligarchy when they become

the majority.



55/60

Conflict over Regimes Different set of issues arise when no smooth transition to

democracy.

Under many plausible scenarios, wealth influences political

power. Consider a reduced-form model of this.

Suppose that the probability that an oligarchy switches to

democracy is qDt = qD (Wt1), where Wt1 = W

Et1 W

Wt1

is the difference between the levels of wealth of the elite and

the citizens at time t 1.

Assume qD () decreasing.



56/60

Conflict over Regimes

Similarly, when democratic, a society becomes oligarchic with

probability

qOt = qO (Wt1)

where now qO () is a non-decreasing function, with qO (0) = 0,so that with perfect equality, there is no danger of switching

back to oligarchy.

Here Wt refers to the income gap between the initial elite

(those with with sj1 = 1) and the citizens.



57/60

Conflict over Regimes (continued)

Then regime dynamics are:

Dt =

0 with probability 1 qD (Wt1) if Dt1 = 0

1 with probability qD (Wt1) if Dt1 = 0

0 with probability qO (Wt1) if Dt1 = 1

1 with probability 1 qO (Wt1) if Dt1 = 1

,

(25)

Suppose that each agent saves out of current income at a

constant (exogenous) rate < 1.

ThereforeWt =

Wt1 + Y

Et1

.



58/60


This implies that in oligarchy

Wt

= t

n=1

nYEt

n(26)

and

limt

Wt = W YE

1 , (27)

where YE

is given by (21).



59/60


This implies that starting from low wealth levels, wealth

inequality will increase in oligarchy.

Therefore, transition from oligarchy to democracy may become

harder as oligarchy persists.

In particular, if there exists W < W such that

qD

W

= 0, then after a certain number of periods, thesociety will be stuck in oligarchy.

In contrast, in democracy, all agents earn the same amount, so

when Dt+k = 1 for all k 0,

Wt+1 = Wt and limt

Wt = 0. (28)

The implication is that a switch back to oligarchy is most likely

soon after a switch from oligarchy to democracy.



60/60

Shortcomings and Future Directions

No managerial delegation. In practice, it may be possible to

remain a member of the elites while delegating control of firms

to high-skill agents.

Entry barriers do not generate direct revenue. If they did, theoligarchy may want to allow some entry while democracy may

also want to use entry barriers.

Develop less reduced-form models of why wealth may influence

political power.

thinking about the rise and decline of nations

Documents