legal essay
313
[ Journal of Law and Economics, vol. 59 (May 2016)]
© 2016 by The University of Chicago. All rights reserved. 0022-2186/2016/5902-0011$10.00
Endogenous Property Rights
Carmine Guerriero University of Bologna
Abstract
Although property rights are key, their determinants are still poorly understood.
When property is fully protected, some potential buyers with valuation higher
than that of original owners are inefficiently excluded from trade because of
transaction costs. When property rights are weak, low-valuation potential buy-
ers inefficiently expropriate original owners. The trade-off between these two
misallocations implies that the protection of property will be stronger the more
heterogeneous the potential buyer’s preferences are. This implication holds true
when part of the population has more political influence on institutional design,
when transaction costs are determined by either market power or asymmetric
information, and when the disincentive effect of weak property rights is taken
into account. Moreover, it is consistent with the relationships between measures
of ethnolinguistic, genetic, and religious diversity and novel data on the rules on
adverse possession and on government takings of real property in 126 jurisdic-
tions.
1. Introduction
Although overwhelming evidence shows that strong property rights foster in- vestment and trade, only recently have economists begun to study the determi- nants of this institution by looking at the trade-off between the dispersed coercive power in a state of anarchy and the predation by a central enforcement authority (Besley and Ghatak 2010). Here I qualify these contributions by incorporating into the economics literature the key insight proposed by a long tradition of legal
I am truly and deeply indebted to Giuseppe Dari-Mattiacci, who contributed to the gathering and coding of the measures of horizontal property rights. In addition, I would like to thank for their gen- erous effort the contributors who provided information about the law in force in their jurisdictions and for their insightful comments an anonymous reader, Gani Aldashev, Kenneth Ayotte, Serra Boranbay, Matteo Cervellati, James D. Dana, Ezra Friedman, Piero Gottardi, Luigi Guiso, Henry Hansmann, Oliver Hart, Richard Holden, Hugo Hopenhayn, Scott Kominers, Clare Leaver, Anup Malani, Corrado Malberti, Giovanni Morelli, Sander Onderstal, Raffaella Paduano, Enrico Perotti, Nicola Persico, Jens Prüfer, Andrea Ruggeri, Larry Samuelson, Kathryn E. Spier, Avraham Tabbach, Kaj Thomsson, Lukáš Tóth, Jeroen van de Ven, Joel Watson, Barry Weingast, Abraham Wickelgren, and participants in seminars at the Einaudi Institute for Economics and Finance, Maastricht Univer- sity, Panthéon-Assas University (Paris II), Tilburg Law and Economics Center, University of Am- sterdam, the 2013 International Society for New Institutional Economics meeting, the 2013 Petralia Sottana Applied Economics Workshop, and 2013 Italian Society of Law and Economics meeting.
314 The Journal of LAW & ECONOMICS
scholarship (Calabresi and Melamed 1972): incomplete property rights can be ef- ficient when transaction costs prevent consensual trade.
In my model, I study a simple yet general economy. Society is equally split into original owners, who have the same valuation for the only good in the economy, and potential buyers, who are instead endowed with heterogeneous valuations. The former are randomly matched to the latter by an intermediation technology that allows each potential buyer to either obtain the good via consensual transfer by paying the original owner’s valuation and bearing a socially wasteful transac- tion cost or to expropriate it at no cost. I define property rights as the probability that an expropriated good is given back to its original owner. When property is fully protected, some potential buyers with valuation higher than that of origi- nal owners are inefficiently excluded from trade because of the transaction cost. When the protection of property is weak, low-valuation potential buyers inef- ficiently expropriate original owners. A rise in the heterogeneity of the poten- tial buyer’s valuations makes inefficient expropriation by low-valuation potential buyers weigh more heavily on social welfare than inefficient exclusion from trade and so induces stronger property rights. This prediction survives if the economy becomes political and thus someone is excluded from the institutional design, when the transaction cost is endogenously determined by either market power or asymmetric information, when potential buyers need to pay damages to keep a successfully expropriated good, when original owners have heterogeneous valua- tions, when individuals can bargain over the price, when expropriation is costly, and when production and investment activities are considered.
The central implication of the model sheds light on two recent cases of wide- spread property rights reforms. First, local politicians in India have been defend- ing the massive postwar condoning of power thefts by the rural population as an attempt to avoid that the homogeneous subsistence farming would be destroyed if electricity invoices, which are mainly constituted by transaction costs, were col- lected (Charnoz and Swain 2012). Second, in the last two decades, the restitution of the lands confiscated by the communist regimes in the former Eastern Bloc has been more limited where farming was more homogeneous, for example, in Alba- nia (Bejtja and Bejtja 2014). In these regions indeed, present-day owners would bear large losses if dispossessed because of the huge transaction costs of acquiring new plots.
Even if anecdotal evidence is instructive, multivariate analysis is more con- vincing. Accordingly, I use a novel data set on the rules on the acquisition of ownership through adverse possession and on the use of government takings to transfer real property from a private party to another private party prevailing in 126 jurisdictions.1 These data measure the extent of protection of property from direct and indirect private takings, which are ubiquitous forms of expropriation that occur daily within the rule of law and are thus different from predation by the state and elites. The latter is much less common but has been the focus of
1 The civil law nomenclature for adverse possession is “acquisitive prescription.”
Endogenous Property Rights 315
the extant literature (Besley and Ghatak 2010).2 Because of this trend, the most widely used proxies for property rights are subjective assessments of the risk of predation by the state and elites (Acemoglu and Johnson 2005), which are thus unsuited to evaluate predictions on legalized takings by generic members of so- ciety.
To capture preference heterogeneity, I focus on the contemporary genetic di-
versity as predicted by the prehistoric migratory distance from East Africa and
post-1500 population flows. This choice is guided by the following two argu-
ments. First, population genetics clarifies that genetic diversity is a primitive
metric of the genealogical distance between populations with a common ancestor
and so of the differences in characteristics transmitted across generations, such as
preferences (Cavalli-Sforza, Menozzi, and Piazza 1994). Consistent with this idea,
Ashraf and Galor (2013) empirically establish that the intensity of contemporary
heterogeneity across genetic, ethnic, and linguistic groups is predominantly de-
termined by genetic diversity within precolonial indigenous settlements. Accord-
ingly, my results are similar when I use either pre colonial genetic diversity or one
among the linguistic, ethnic, or religious fractionalizations. Second, the prehis-
toric exodus of Homo sapiens out of Africa occurred in a series of stages such that
new groups carried from initial colonies only a portion of their genetic heteroge-
neity, and thus precolonial genetic diversity is strongly negatively correlated with
prehistoric migratory distance from Addis Ababa.
Ordinary least squares —OLS—estimates suggest that the protection of the
original owner’s property rights is the strongest where the extent of contempo-
rary genetic diversity is the largest. This evidence is consistent with the model,
and because of the primarily prehistoric origins of contemporary genetic pools,
it is quite unlikely that it is driven by reverse causation. I cannot, however, ex-
clude that these results are produced by the confounding impact of the Europe-
ans’ colonization strategy on contemporary genetic diversity and the design of
individual rights and/or by other determinants of property rights associated with
preference diversity.
Building on this remark, I pursue a number of strategies to determine if the
correlations I uncover are, in fact, causal. First, I control for the three key deter-
minants of the colonizers’ settlement strategy—the deadliness of the disease en-
vironment, precolonial prosperity, and the duration of colonization (Acemoglu,
Johnson, and Robinson 2002)—and for the development level, the inclusiveness
of political institutions, external and internal conflicts, other drivers of state ca-
pacity such as legal origins, and geographic features affecting preference hetero-
geneity and the structure of the economy. Considering these observables either
stepwise or together leaves the results almost intact. Second, I show that the es-
timates are similar when I focus on those jurisdictions in which the majority of
2 Some of these odd cases are reviewed by Besley and Ghatak (2010). In contrast, the sectors affected by private takings account for 32 percent of the 2011 US gross domestic product (Dari- Mattiacci, Guerriero, and Huang 2016).
316 The Journal of LAW & ECONOMICS
contemporary inhabitants can trace their ancestry back to 1500 to the same area.
Together, these two exercises exclude that colonial rule is driving the patterns I
uncover in the data. Third, I use recent insights from Altonji, Elder, and Taber
(2005) to calculate how much greater the influence of unobservable factors, rela-
tive to observable factors, would need to be to explain away the positive relation-
ship between contemporary genetic diversity and property rights. I find that the
influence of unobservables would have to be on average more than twelve times
greater than that of all observables considered together. Given the high fit of the
regressions, this is unlikely. Finally, I exploit the plausible exogeneity of the pre-
historic migratory distance from East Africa and I use it as an excluded instru-
ment for contemporary genetic diversity. The two-stage least squares—2SLS—
estimates are qualitatively identical to the OLS estimates, so I cannot reject that
contemporary genetic diversity is exogenous at a level nowhere lower than 5 per-
cent, conditional on all observables. These robustness checks make it difficult to
envision that OLS correlations are driven by unobservables and, in particular, a
mechanism different from the one I model. Hence, I take them as consistent with,
if not proving, causality running from preference diversity to property rights. In
addition, my private-taking-based measures of property rights correlate weakly
and possibly negatively with the most used predation-based proxies for property
rights. Therefore, a key avenue for future research is to study the relationships
among the horizontal property rights I study while testing my model, the vertical
property rights defending the citizenry from predation by the state and elites and
analyzed in the literature, and the economy. The present study is related to four bodies of research. The first one looks at the
increasing importance of human capital as compared to natural resources and, in general, at the falling income inequality as a driver of efficient political insti- tutions safeguarding individuals from expropriation (Cervellati, Fortunato, and Sunde 2008). In my model instead, a smaller diversity in the preferences and/or abilities of a group of outsiders weakens the property rights of a group of insiders who initially hold all the value in society. A second line of papers focuses on the relative importance of protecting property rights and enforcing contracts (Ace- moglu and Johnson 2005), whereas I examine the determinants of the trade-off between these two institutional strategies created by the possibility of transfer- ring value without consent (see also Dari-Mattiacci, Guerriero, and Huang 2016). Hence, the present paper also contributes to the extensive literature comparing property rules and liability rules (Calabresi and Melamed 1972). A third group of studies shows that weak property rights can be optimal in an endowment econ- omy (Jordan 2006; Piccione and Rubinstein 2007; Bar-Gill and Persico, forth- coming). Not only do I extend this result to production economies, but I also stress the pivotal role of preference heterogeneity. Finally, a lively literature sees the legal system as an optimal response to the risk that the majority of agents are coerced by more powerful special interests (Glaeser and Shleifer 2003), by a mi- nority of similarly powerful untrustworthy agents (Aghion et al. 2010), or by the state (Besley and Ghatak 2010). I broaden the scope of this literature by consid-
Endogenous Property Rights 317
ering the problem of legal direct and indirect nonconsensual transfers between private parties.
The paper proceeds as follows. In Section 2, I outline the basic relationship be- tween horizontal property rights and contemporary genetic diversity to motivate the general model of endogenous property rights that I discuss in Section 3. I as- sess how robust the model’s prediction is to alternative assumptions in Section 4 and whether the motivating correlations are indeed causal in Section 5. Finally, I conclude in Section 6. The list of countries in my sample and formal proofs of the model’s results are in Appendixes A and B, respectively.
2. Horizontal Property Rights and Genetic Diversity
While all legal systems punish theft and provide remedies for dispossessed owners, the protection of property against private expropriation is almost never complete. When transaction costs are sizable indeed, private parties are often al- lowed to take private property with or without paying compensation to the origi- nal owner (Bouckaert and De Geest 1995). There are two main instances in which this legal private expropriation takes place.
The first one is adverse possession of personal and real property, that is, mov- able and immovable goods in civil law parlance. Adverse possession is a form of acquisition of property rights such that a possessor becomes the legal owner of a good by virtue of the passage of time and recurring other conditions such as open, continuous, and notorious possession. By means of adverse possession, property can be effectively legally transferred without the original owner’s consent. Dif- ferent legal systems balance the interests of the original owner and the possessor differently. In the United States, adverse possession of personal property is not admitted, whereas a good-faith possessor of a movable acquires ownership after 10 years in Germany (BGB 937) and 3 years in France (C. Civ. 2276). Accord- ingly, an original owner is protected the most in the United States and more in Germany than in France. Hence, a natural measure of an original owner’s prop- erty rights on a movable good is the number of years after which a good-faith possessor of the good acquires ownership or thirty years, which is the maximum length of adverse possession observed in the sample, if the good-faith posses- sor cannot acquire property rights (Dari-Mattiacci and Guerriero 2015). To con- struct this measure, AP-Movable, and the other measures of property rights de- scribed below, Giuseppe Dari-Mattiacci and I sent questionnaires to members of Lex Mundi and HG.org (see Djankov et al. 2003), participants in the World Bank Doing Business project, and law scholars. These contributors, whose names and affiliations are listed in the online appendix, based their answers on the laws and judicial decisions that prevailed in their jurisdiction between 1981 and 2011. The lack of any reform leaves me with a cross section of 126 jurisdictions (see Appen- dix A). Table 1 reports the sources and summary statistics for my variables.
Adverse possession of real property is more involved, since the transfer of im- movables requires some form of registration, which has different effects across legal systems. Hence, I focus on AP-Immovable, which is defined as the num-
318
T a
b le
1
S u
m m
a r
y o
f V
a ri
a b
le s
V a ri
a b
le D
efi n
it io
n S
a m
p le
S
iz e
M ea
n S
o u
rc e
H o
ri zo
n ta
l p
ro p
er ty
r ig
h ts
:
A
P -M
o v
a b
le Y
ea rs
n ee
d ed
f o
r a d v
er se
p o
ss es
si o
n b
y a
n y
g o
o d
-f a it
h p
o ss
es so
r o
f a
m o v
a b
le g
o o
d 1
2 6
1 1
.3 0
6 (1
1 .8
9 0
) D
a ri
-M a tt
ia cc
i a n
d G
u er
ri er
o (
2 0
1 6
)
A
P -I
m m
o v
a b
le Y
ea rs
n ee
d ed
f o
r a d v
er se
p o
ss es
si o
n w
it h
o u
t ti
tl e
b y
a g
o o
d -f
a it
h p
o ss
es so
r o
f a n
i m
m o v
a b
le g
o o
d 8
9 2
4 .5
7 3
(8
.0 8
2 )
D a ri
-M a tt
ia cc
i a n
d G
u er
ri er
o (
2 0
1 6
)
G
o v
er n
m en
t- T
a k
in g
s D
u m
m y
v a ri
a b
le e
q u
a l
to o
n e
if t
h e
st a te
c a n
t a k
e re
a l
p ro
p er
ty f
ro m
a
p ri v
a te
p a rt
y o
n ly
f o
r p
u b
li c
u se
b u
t n
o t
to t
ra n
sf er
i t
to a
n o
th er
p ri v
a te
p
a rt
y f
o r
p ri v
a te
f o
r- p
ro fi
t u
se ,
ze ro
o th
er w
is e
8 8
.7 1
6 (.
4 5
4 )
D a ri
-M a tt
ia cc
i a n
d G
u er
ri er
o (
2 0
1 6
)
V er
ti ca
l p
ro p
er ty
r ig
h ts
:
P
ri v
a te
-P ro
p er
ty F
iv e-
ca te
g o
ry s
co re
t h
a t
in cr
ea se
s w
it h
p er
ce iv
ed p
ro te
ct io
n o
f p
ri v
a te
p
ro p
er ty
f ro
m g
o v
er n
m en
t ex
p ro
p ri
a ti
o n
i n
1 9
9 7
1 1
3 3
.4 4
2 (1
.1 2
6 )
G w
a rt
n ey
a n
d L
a w
so n
( 1
9 9
7 )
A
v -E
x p
ro p
ri a ti
o n
T en
-c a te
g o
ry i
n d
ex w
it h
h ig
h er
v a lu
es w
h en
t h
e p
er ce
iv ed
r is
k o
f ex
p ro
p ri
a ti
o n
o f
p ri v
a te
f o
re ig
n i
n v
es tm
en t
b y
t h
e g
o v
er n
m en
t a n
d
p o
w er
fu l
el it
es w
a s
lo w
er b
et w
ee n
1 9
8 5
a n
d 1
9 9
5
9 7
7 .3
7 6
(1 .6
7 0
) P
o li
ti ca
l R
is k
S er
v ic
es (
h tt
p s:
// w
w w
. p
rs g
ro u
p .c
o m
/)
P re
fe re
n ce
h et
er o
g en
ei ty
:
G
en et
ic -D
iv er
si ty
-A A
n ce
st ry
-a d
ju st
ed e
x p
ec te
d h
et er
o zy
g o
si ty
, a s
p re
d ic
te d
b y
t h
e p
re h
is to
ri c
m ig
ra to
ry d
is ta
n ce
f ro
m E
a st
A fr
ic a a
n d
p o
st -1
5 0
0 p
o p
u la
ti o
n fl
o w
s 1
2 6
.7 2
5 (.
0 2
7 )
A sh
ra f
a n
d G
a lo
r (2
0 1
3 )
G
en et
ic -D
iv er
si ty
E x
p ec
te d
h et
er o
zy g
o si
ty ,
a s
p re
d ic
te d
b y
t h
e p
re h
is to
ri c
m ig
ra to
ry
d is
ta n
ce f
ro m
E a st
A fr
ic a
1 2
6 .7
0 8
(. 0
5 3
) A
sh ra
f a n
d G
a lo
r (2
0 1
3 )
L
a n
g u
a g
e N
a tu
ra l
lo g
a ri
th m
o f
th e
n u
m b
er o
f la
n g
u a g
es 1
1 5
2 .3
6 7
(1 .5
2 3
) M
ic h
a lo
p o
u lo
s (2
0 1
2 )
E
th n
ic E
th n
ic f
ra ct
io n
a li
za ti
o n
1 2
5 .4
0 6
(. 2
4 2
) A
le si
n a e
t a l.
( 2
0 0
3 )
R
el ig
io n
R el
ig io
u s
fr a ct
io n
a li
za ti
o n
1 2
5 .4
4 5
(. 2
2 4
) A
le si
n a e
t a l.
( 2
0 0
3 )
O th
er c
o n
tr o
ls :
P
a th
o g
en -L
o a d
M ea
su re
i n
z -s
co re
s o
f th
e h
is to
ri ca
l p
re v
a le
n ce
o f
se v
en d
is ea
se -c
a u
si n
g
p a th
o g
en s
(l ei
sh m
a n
ia ,
tr y
p a n
o so
m a ,
m a la
ri a ,
sc h
is to
so m
a ,
fi la
ri a ,
sp ir
o ch
et es
, a n
d l
ep ro
sy )
1 2
3 .1
0 1
(. 6
2 8
) M
u rr
a y
a n
d S
ch a ll
er (
2 0
1 0
)
P
o p
u la
ti o
n -D
-1 5
0 0
N a tu
ra l
lo g
a ri
th m
o f
p o
p u
la ti
o n
d en
si ty
, d
efi n
ed a
s th
e ra
ti o
o f
th e
to ta
l p
o p
u la
ti o
n t
o a
ra b
le l
a n
d ,
in 1
5 0
0 1
2 2
1 .1
0 5
(1 .5
7 6
) M
cE v
ed y
a n
d J
o n
es (
1 9
7 8
)
319
C
o lo
n iz
a ti
o n
-T im
e D
u ra
ti o
n o
f th
e co
lo n
iz a ti
o n
e x
p er
ie n
ce i
n h
u n
d re
d s
o f
y ea
rs 1
1 3
1 .1
3 5
(1 .7
4 3
) A
sh ra
f a n
d G
a lo
r (2
0 1
3 )
P
o p
u la
ti o
n N
a tu
ra l
lo g
a ri
th m
o f
th e
2 0
1 0
t o
ta l
p o
p u
la ti
o n
1 2
6 1
6 .3
7 5
(1 .5
4 4
) O
E C
D ,
P o
p u
la ti
o n
( h
tt p
:/ /s
ta ts
.o ec
d .
o rg
/I n
d ex
.a sp
x ?D
a ta
se tC
o d
e= P
O P
_
F IV
E _
H IS
T )
In
co m
e N
a tu
ra l
lo g
a ri
th m
o f
2 0
1 0
r ea
l g
ro ss
d o
m es
ti c
p ro
d u
ct p
er c
a p
it a r
el a ti v
e to
t h
e U
n it
ed S
ta te
s a t
2 0
1 0
p ri
ce s
1 2
4 2
.9 7
1 (1
.2 8
7 )
P en
n W
o rl
d T
a b
le 7
.1 (
h tt
p s:
// p
w t.
sa s
.u p
en n
.e d
u /)
.
N
eo li
th ic
A v
er a g
e ti
m e
si n
ce a
g ri
cu lt
u ra
l tr
a n
si ti
o n
i n
t h
o u
sa n
d s
o f
y ea
rs 1
1 9
5 .2
3 3
(2 .3
6 1
) P
in h
a si
, F
o rt
, a n
d A
m m
er m
a n
( 2
0 0
5 )
D
em o
cr a cy
C o
n st
ra in
ts o
n t
h e
ex ec
u ti v
e a u
th o
ri ty
s co
re r
a n
g in
g b
et w
ee n
1 a
n d
7 a
n d
a v
er a g
ed o
v er
a ll
a v
a il
a b
le y
ea rs
b ef
o re
2 0
1 0
1 1
9 4
.1 4
9 (1
.8 2
8 )
P O
L IT
Y I
V 2
0 1
1 V
er si
o n
( h
tt p
:/ /
w w
w .s
y st
em ic
p ea
ce .o
rg )
R
es er
v es
V a lu
e o
f 2
0 1
0 c
ru d
e o
il p
ro v
ed r
es er
v es
p er
c a p
it a i
n 2
0 1
0 U
S d
o ll
a rs
1 2
6 4
.3 E
7 (2
.3 E
8 )
E n
er g
y I
n fo
rm a ti
o n
A d
m in
is tr
a ti
o n
(h
tt p
s: //
w w
w .e
ia .g
o v
/c fa
p p
s /i
p d
b p
ro je
ct /i
ed in
d ex
3 .c
fm ?t
id =
5
& p
id =
5 7
& a id
= 6
& ci
d =
re g
io n
s &
sy id
= 2
0 0
9 &
ey id
= 2
0 1
1 &
u n
it =
B B
)
C
o n
fl ic
t- E
S h
a re
o f
y ea
rs b
et w
ee n
1 8
2 3
a n
d 2
0 0
7 i
n w
h ic
h t
h e
ju ri
sd ic
ti o
n w
a s
in v
o lv
ed i
n a
n e
x te
rn a l
m il
it a ry
c o
n fl
ic t
1 2
6 .0
1 9
(. 0
3 2
) C
o rr
el a te
s o
f W
a r
(h tt
p :/
/w w
w
.c o
rr el
a te
so fw
a r.
o rg
/d a ta
-s et
s /f
o ld
er _
li st
in g
)
C
o m
m o
n -L
a w
D ec
en tr
a li
za ti
o n
o f
th e
ju ri
sd ic
ti o
n ’s
l eg
a l
tr a d
it io
n i
n 2
0 0
0 r
a n
g in
g
b et
w ee
n 0
a n
d 1
a n
d a
ss u
m in
g h
ig h
er v
a lu
es t
h e
n ea
re r
it i
s to
a
p er
fe ct
ly d
ec en
tr a li
ze d
c o
m m
o n
-l a w
s y
st em
9 3
.3 8
6 (.
2 7
7 )
G u
er ri
er o
( 2
0 1
6 )
C
o n
fl ic
t- I
S h
a re
o f
th e
y ea
rs b
et w
ee n
1 8
1 6
a n
d 2
0 0
7 i
n w
h ic
h t
h e
ju ri
sd ic
ti o
n w
a s
in v
o lv
ed i
n a
n i
n te
rn a l
m il
it a ry
c o
n fl
ic t
1 2
6 .0
3 1
(. 0
6 4
) C
o rr
el a te
s o
f W
a r
(h tt
p :/
/w w
w
.c o
rr el
a te
so fw
a r.
o rg
/d a ta
-s et
s /f
o ld
er _
li st
in g
)
L
a n
d -Q
u a li
ty -S
D S
ta n
d a rd
d ev
ia ti
o n
o f
la n
d q
u a li
ty ,
w h
ic h
r a n
g es
b et
w ee
n 0
a n
d 1
a n
d i
s ca
lc u
la te
d a
s th
e p
ro d
u ct
o f
a m
ea su
re o
f th
e cl
im a ti
c su
it a b
il it
y f
o r
cu lt
iv a ti
o n
a n
d a
m ea
su re
o f
th e
so il
s u
it a b
il it
y f
o r
cu lt
iv a ti
o n
1 1
5 .2
0 2
(. 0
9 3
) M
ic h
a lo
p o
u lo
s (2
0 1
2 )
R
u g
g ed
n es
s R
u g
g ed
n es
s o
f th
e te
rr a in
i n
k il
o m
et er
s 1
2 3
.1 8
4 (.
1 2
3 )
G -E
co n
( h
tt p
:/ /g
ec o
n .y
a le
.e d
u /d
a ta
-a
n d
-d o
cu m
en ta
ti o
n -g
-e co
n -p
ro je
ct )
L
a ti
tu d
e A
b so
lu te
l a ti
tu d
e n
o rm
a li
ze d
t o
r a n
g e
b et
w ee
n 0
a n
d 1
1 2
6 .3
3 2
(. 1
9 1
) W
o rl
d C
IA F
a ct
b o
o k
( h
tt p
s: //
w w
w .c
ia
.g o
v /)
E x
cl u
d ed
i n
st ru
m en
t:
M
ig ra
to ry
-D is
ta n
ce P
re h
is to
ri c
m ig
ra to
ry d
is ta
n ce
f ro
m A
d d
is A
b a b
a i
n t
h o
u sa
n d
s o
f k
il o
m et
er s
1 2
6 8
.9 8
1 (6
.9 6
8 )
A sh
ra f
a n
d G
a lo
r (2
0 1
3 )
N o
te .
S ta
ti st
ic s
a re
c o
m p
u te
d f
o r
th e
m a x
im u
m a
v a il
a b
le s
a m
p le
.
320 The Journal of LAW & ECONOMICS
ber of years necessary for adverse possession of immovable goods with good faith but without title if the adverse possessor prevails against a good-faith buyer from the original owner or if registration is compulsory for adverse possession and thirty years if either registration is not compulsory for adverse possession and the good-faith buyer from the original owner prevails over the adverse pos- sessor or if the adverse possessor’s ownership is not legally recognized.3 Unlike AP-Movable, AP- Immovable is constructed with a particular scenario in mind: the conflict between a good-faith possessor of real property without title and a third party buying from the original owner after the term for adverse possession has expired. This legal case most emphasizes the adverse possessor’s claim since enforceability against third parties is the key feature of property rights on im- movables ( Bouckaert and De Geest 1995). When the possessor prevails, the legal protection of the original owner and, in turn, of the buyer is weaker.
The second instance of legal private expropriation concerns takings that are indirect and, in particular, mediated by the state (Bell 2009). Even if all jurisdic- tions allow the state to seize private real property under the power of eminent do- main for public interest and after paying just compensation (Fischel 1995), they differ in the extent to which the transfer of real property from a private party to another qualifies as being in the public interest. Hence, I consider the dummy variable Government-Takings, which equals one if the state can take real prop- erty from a private party only for public use but not transfer it to another private party for private for-profit use. Accordingly, when Government-Takings is zero, the state can use government takings to legally favor a subgroup of the popula- tion to the detriment of another.
Figures 1A, B, and C show the sizable variation in the three measures of hori- zontal property rights over the entire sample. To illustrate, AP-Movable displays a mean value of 11.306 years and a standard deviation of 11.890 years (see Ta- ble 1), which captures the existence of a variety of legal solutions ranging from full protection of the original owner in the United States to full protection of the good-faith possessor in Italy. To construct the maps, I divide the range of each variable into four equal intervals. In the empirical test I instead use continuously measured variables.
Following Ashraf and Galor (2013), I construct a measure of contemporary ge- netic diversity, which, as discussed in Section 1, is my proxy for preference di- versity, using the index of aggregate expected heterozygosity in the precolonial indigenous settlement of each ethnic group in the jurisdiction. This can be in- terpreted as the probability that two randomly selected individuals in the rele- vant ancestral population differed genetically from one another with respect to a given spectrum of traits,4 and it can be calculated for 53 ethnic groups building on data on allele frequencies from the Human Genome Diversity Cell Line Panel.
3 I assume that a possessor registers the good if necessary to acquire property via adverse posses- sion.
4 To elaborate, the aggregate expected heterozygosity Hexp equals 2
111 (1/ ) , l
i
km ilm p where pi
denotes the frequency of the ith allele among the kl observed variants for locus on gene l and m is the number of possible loci.
Figure 1. Property rights and genetic diversity. A, AP-Movable: B, AP-Immovable; C, Government- Takings; D, Genetic-Diversity-A.
A
B
C
D
322 The Journal of LAW & ECONOMICS
Because property rights are measured at the jurisdiction level, however, I use the indices predicted by Ashraf and Galor (2013) through the serial founder effect— the highly significant negative relationship between precolonial genetic diversity and the prehistoric migratory distance from Addis Ababa—for the ancestral pop- ulations of a cross section including the 126 jurisdictions in my sample. All in all, the proxy variable Genetic-Diversity-A ranges between 0 and 1 and incorporates the average aggregate expected heterozygosity within the ancestral population of each ethnic group in the jurisdiction weighted by post-1500 population flows and the average genetic distance between the jurisdiction’s ancestral population and the ancestral populations of the ethnic groups that immigrated after coloniza- tion (Ashraf and Galor 2013). This choice is validated by two facts: First, grow- ing empirical evidence highlights the relationship between genetic and in general cultural diversity and preference heterogeneity.5 Second, my results are similar when I use either the genetic diversity within the jurisdiction’s ancestral popula- tion or one among the linguistic, ethnic, or religious fractionalizations (Section 5.1.5). Figure 1D displays the sizable variation in Genetic-Diversity-A over the sample.
A comparison among the four maps in Figure 1 illustrates the strong posi- tive correlation between legal protection of the original owner’s property from direct and indirect private takings and preference heterogeneity as captured by the contemporary genetic diversity. Column 1 of Table 2 reports estimates of this correlation from regressions obtained, for ease of comparison, from the subsamples for which I observe all the additional regressors I consider below. A one-standard- deviation rise in Genetic-Diversity-A—.03—is associated with an increase of 3.43 (2.15) years in AP-Movable (AP-Immovable) and with a .12 rise in the expected value of Government-Takings. These coefficients are significant at 10 percent or better.
Next I present a model of the design of property rights in a heterogeneous soci- ety that rationalizes the evidence discussed so far but applicable to a much wider range of legal cases.
3. Theory
3.1. Model’s Setup
3.1.1. Preferences
I consider a society composed of a mass 1 of original owners and a mass 1 of potential buyers, all having linear utility over a good x (see note 12 for the risk- aversion case). While the original owners value x at v, Δ/2 potential buyers value it at 0, Δ/2 of them value x at ,v and the remainder have a valu-
5 While Desmet et al. (2011) document the strong connection between pre-1500 genetic dis- tance and a measure of preference differences based on answers to World Value Survey questions about life, family, religion, and morals in Europe, Lieberman and McClendon (2013) exploit Afro- barometer survey data from 18 sub-Saharan countries to show that ethnicity shapes preferences and, in particular, the propensity to commit crimes and how salient respondents perceive educa- tion, electricity, farming, health, and unemployment policies to be.
Endogenous Property Rights 323
ation λ uniformly distributed over [ , ] with l and m ( )/2. Thus, a rise in � ∈ (0, 1) implies a mean-preserving spread of the λ distribution (see Aghion, Alesina, and Trebbi 2004). I employ the uniform distribution to ob- tain closed-form solutions, but in Appendix B I show that the model’s message holds when I consider a generic probability density function.
3.1.2. Market Structure
To capture institutions of the type discussed in Section 2, I examine a simple yet general market design. Original owners are randomly matched to potential buyers by an intermediation technology that allows each potential buyer to either obtain x with the original owner’s consent by paying v to him and bearing a pos- itive transaction cost α or expropriate it at no cost.6 The expense α has no social value and gauges inefficiencies like the costs borne by potential buyers to borrow v, bargaining costs, or the markup imposed by a foreign intermediary. I posit that α < min{v, l - v} and I discuss the role of this assumption in notes 9, 10, and 11. An expropriated x is returned to its original owner with probability γ, which thus summarizes the legal protection of the original owner’s property rights. In other words, the original owner is protected by a property rule with probability γ, and the potential buyer can keep the good otherwise, that is, when protected by a liability rule. In Section 4.2.1, I explicitly analyze the case of positive damages linking, at the same time, the present model to the literature on property rules versus liability rules.
One glaring example of the type of expropriation I study is that of an inter- mediary (agent) selling to a good-faith buyer a good stolen (embezzled) from an original owner (his principal) at a low price. As happens in the model, the good is then possibly given back to its original owner, and the strength of property rights reduces the probability that the buyer consumes. To elaborate, 1 - γ will be lower the stronger the remedies in the original owner’s hands are, the longer the buyer needs to wait before acquiring ownership by adverse possession, and the more effective public enforcement is (Dari-Mattiacci, Guerriero, and Huang 2016). The model applies, however, to a wider range of cases. First, expropriation can assume the form of squatting on a plot or in a house. Second, the good can be envisaged as an input producing a fixed value when transformed via the old tech- nology in the hands of the original owners and an uncertain one when the new technology is applied by the potential buyers. Third, the conflict between an orig- inal owner and a potential buyer can be reinterpreted as the one involving a cred- itor and a shareholder who tunnels resources out of a firm (Johnson et al. 2000), that arising between a buyer and a seller who breaches the contract, or that possi- bly setting an agent against a principal in a segmented labor or financial market. Finally, if the taking is indirect, the good can be conceived of as having a fixed
6 More generally, this intermediation technology can be also seen as an efficient market maker that assigns high-valuation potential buyers to original owners in exchange for the wasteful pay- ment α.
324 The Journal of LAW & ECONOMICS
value for the original owner and an uncertain one for the other private party to which the state transfers it after having decided to expropriate instead of buying.7
3.1.3. Timing
At time t0, γ is chosen to maximize social welfare, which is the sum of the orig- inal owner’s and potential buyer’s utilities. At t1, individuals learn who they are and, thus, their valuation. At t2, individuals are matched by the intermediation technology. Finally at t3, any expropriated x is returned to its original owner with probability γ.
3.1.4. Interpretation
In interpreting the generality of the foregoing, several points should be stressed. First, an alternative instrument that society has is to decrease α. This policy re- duces market inefficiencies and possibly leads to complete property rights (see notes 9 and 10), but it is very costly, if not impossible, to implement because of political imperfections (Besley and Ghatak 2010) and the opposition of agents who gain from large transaction costs (see Section 4.1.1 and Barry, Hatfield, and Kominers 2014). Second, the model’s prediction survives when the economy be- comes political and thus someone is excluded from the institutional design (see Section 3.3). Third, it is unaffected when the transaction cost is endogenously de- termined by either market power or asymmetric information (see Section 4.1). Fourth, it holds when original owners have heterogeneous valuations, individ- uals can bargain over the price, and expropriation is costly (see Section 4.2). Fi- nally, it is robust to the introduction of production and investment activities (see Section 4.2).
3.2. Socially Optimal Property Rights
A potential buyer buys if her valuation λ net of the purchasing costs v + α is greater than her expected payoff from expropriation (1 - γ)λ, which is the case if ( .ˆ )/v 8 When selecting γ*, where the asterisk denotes an equilib- rium choice, society then maximizes
ˆ
ˆ
( ) (1 ) 2
(1 ) (1 ) (1 )
2
d l
v v d
l
(1)
7 This last interpretation can be fruitfully applied to study partial taking law by further assuming that after partial expropriation the state pays severance damages (1 - γ)v to the original owner to compensate him for the loss (Sackman et al. 2013). Then the equilibrium is second best (see the case in which ϕ = 0 in Section 4.2.1).
8 Provided that only one good can be consumed, the results continue to hold even if the goods that are purchased can be expropriated before consumption since all those willing to expropriate have already done so.
Endogenous Property Rights 325
for ˆ or * (( )/ , 1]v and social welfare under full expropriation
FE
m m
(1 ) [(1 ) ] (1 ) (1 )
v v d v W
l
otherwise.9 Switching from complete to incomplete property rights—that is, from γ* = 1 to γ* < 1—has three effects: it saves α at the cost of misallocating x with probability γ* for the ˆv matches, it avoids misallocation with prob- ability 1 - γ* for the v ≤ λ < v + α matches by expanding the consumption set of the potential buyers, and it misallocates x with probability 1 - γ* for the λ < v matches. While this last effect is negative, the sum of the first two is positive, provided that α is sufficiently large.10 For ˆ , the solution to society’s prob- lem is uniquely defined by the necessary and sufficient first-order condition
2 2
1 2 2 2
ˆ ( )ˆ ˆ ˆ2 ( * * ) ( ) 2 ( ) 0 1
( *) ( ) ( ) . 1
d v l v v
d
lv v v
(2)
The first line of equation (2) implies that a rise in γ has a marginal effect that is positive and an inframarginal effect
ˆ 1 ˆ ˆ(1 ) ( ) ( 2 ( ) ,
2 2 2
v d v v v
l l
which can be negative only if v < λm. For v ≥ λm then, optimal property rights are complete since W FE also rises with γ. For v < λm instead, W
FE decreases with γ and γ* ≤ 1 is possible. It equals either the level implicitly defined by the second line of equation (2) or 0 depending on which of the two maximizes social welfare. The former case is more likely the larger � is under the restrictions on α (see Fig- ure 2).11 Intuitively, if � grows bigger, the welfare returns from the expropriation activities of the potential buyers for which ˆv fall and the welfare costs from the expropriation by low-valuation potential buyers rise. Thus, γ* has to in- crease. The role of the aggregate uncertainty l is similar; γ* also weakly rises with the original owner’s valuation since an increase in v has the inframarginal effect of boosting the original owner’s payoff when his property rights are protected and the marginal effect of raising ˆ: both effects call for a higher γ*. Finally, γ*
9 The assumption α < v assures that problem (1) is concave. Should it fail, γ* will be 1 (0) for v < λm if social welfare is greater at 1 (0) than it is at 0 (1) or
2 ( ) (1 )[v l
2 m( ) 2( )] ( ) 2 ,v v l which can be the case for Δ → 1 (0). The inequality leading
to complete property rights is easier to satisfy the larger Δ is if ( )( )v v or ,v which is true under my hypotheses.
10 To be precise, if α2 > (1 - γ)v2. For Δ = 0, γ* < 1 whenever ,v which is possible given my assumptions.
11 The exact condition is 2 22 ( * ) ( )/ * * 2 ( ),m v v v v which is true when α < l - v ( v ) if it is easier to satisfy the larger (smaller) γ* is. Both inequalities hold under my hypotheses. For α > λm - v, the above condition is more difficult to satisfy the larger α is, as in Figure 2, where .
326 The Journal of LAW & ECONOMICS
decreases with α if interior (see Figure 2). Proposition 1 summarizes the analysis of this section:12
Proposition 1. The optimal level of property rights γ* rises weakly with the original owner’s valuation v, the aggregate uncertainty about the potential buy- er’s valuation l, and preference heterogeneity Δ. In addition, it decreases with the transaction cost α if interior.
These results are strongly related to the out-turn of several literatures and con- sistent with substantial anecdotal evidence on the determinants of institutional variation. First, they are linked to models that predict a positive relationship be- tween defensive patenting and the dispersion of the technological base on which a new product builds (see Hall and Harhoff 2012). Second, they are in the same spirit of findings linking taxation to preference diversity (Saez 2002; Diamond and Spinnewijn 2012). It is important to notice, however, that the tax γ only sometimes redistributes value from high- to low-valuation individuals. Third, the model’s main result yields important insights for empirical analyses of the theory of the firm. Greater transaction costs call for more vertical integration because of asset specificities (Williamson 2010), but they also shape the upstream suppliers’ property rights and, therefore, the risk of holdup (Hart and Moore 1990). Thus, the direct and indirect effects of transaction costs should be considered when an- alyzing the boundaries of the firm.
Finally, the model’s key result is not only consistent with the evidence dis- cussed in Section 2, but it also sheds light on two recent cases of widespread property rights reforms. First, it rationalizes the postwar condoning of thefts of
12 Following Aghion, Alesina, and Trebbi (2004), I can introduce risk aversion into the setup by assuming that individuals who gain an expected utility lower than that prevailing under the certain scenario of full property rights also incur a loss u. Since all original owners (potential buyers) weakly prefer complete (incomplete) property rights, a rise in risk aversion is isomorphic to a fall in v and thus induces a weakly lower γ*.
Figure 2. Property rights, preference heterogeneity, and transaction costs. The dotted line depicts the relationship between property rights and preference heterogeneity for a level of transaction costs higher than that producing the solid line.
Endogenous Property Rights 327
power by Indian farmers. In fact, despite the fact that the annual cost of this pol- icy is around 1.5 points of the 2012 gross domestic product (GDP), local pol- iticians have strenuously defended it by asserting that collecting the electricity invoices, which are mainly constituted by billing costs, would destroy homoge- neous subsistence farming (Charnoz and Swain 2012). Second, the model clar- ifies the asymmetric restitution of the lands confiscated by communist regimes in the former Eastern Bloc. The original owner’s property rights have been least protected where farming was less productive and thus more homogeneous, for example, in Albania (Bejtja and Bejtja 2014). There, present-day owners would bear large losses if dispossessed because of the huge transaction costs of acquiring new plots.
3.3. The Political Economy of Property Rights Protection
Thus far I have examined the case involving a perfect veil of ignorance behind which everybody is identical. Yet reality is often different. To evaluate the positive side of the design of property rights, I follow Alesina, Aghion, and Trebbi (2004) and I consider a situation in which the minority of individuals who choose γ also knows their future role in the economy and can exclude the rest of the population from maximization of the social welfare function. It seems natural to think that these insiders are the original owners and potential buyers with the highest valu- ation—that is, whose [ , ]. Accordingly, γ* now maximizes
ˆ
ˆ
(1 ) ( ) (1 ) (1 )
2 2
v d d v
l l l
for ˆ and
2
FE 1 2 (1 )
2 W
l
otherwise. Comparing both expressions with equation (1) suggests that equilib- rium property rights in a political economy still rise with � for ε not too small, but they are set inefficiently high by individuals who do not face uncertainty (see Appendix B); for example, the Zamindari system of taxation allowed Indian land- owners to expropriate evading tenants who were often more productive (Besley and Ghatak 2010). If instead the excluded potential buyers have valuations higher than , equilibrium property rights rise weakly with � for ε not too small
and equal (higher than) the γ* found in the basic setup if interior (otherwise), as illustrated in Appendix B.
4. Robustness to Alternative Assumptions
Next, I document the robustness of the model’s message to alternative hypoth- eses.
328 The Journal of LAW & ECONOMICS
4.1. Endogenous Transaction Costs
4.1.1. More General Market Structures
Consider the case in which the intermediation technology is in the hands of the original owners and the transaction costs equal α = βδ, where δ is an in- verse measure of market competitiveness and β is the markup the original own- ers apply to v and select between t1 and t2. In particular, β* maximizes the sum of the original owner’s expected profits and the expected payoff from consuming x when returned or
ˆ ˆ( )( ) ( )
( ) (1 ) (1 ) * * 2 2
v v v v
l l
for ) * ˆ( /v and γ*v otherwise. Markup β* can be positive only in the former case when it equals
* * (2 *)
* 4(1 ) 2 2
l v
and thus decreases with both the degree of market competitiveness δ and the probability of legal expropriation 1 - γ*. Hence, stronger property rights con- temporaneously decrease a potential buyer’s payoff from expropriating and, be- cause of the rise in β*, her utility from buying x. The optimality of the original owner’s choice and the uniformity of preferences imply that these two effects cancel out, so *ˆ/ 0d d . Since α is now a transfer, γ* < 1 only when the distortions of the potential buyer’s demand due to market power are sizable. To elaborate, optimal property rights maximize
ˆ
ˆ
(1 ) (1 ) (1 ) (1 )
2 2
v v d d
l l
for ˆ and W FE otherwise. For v < λm, the derivative of the social welfare function with respect to γ is negative for ˆ and concave in Δ, positive for Δ → 1, and possibly negative for Δ → 0 if ˆ . Then γ* can again jump from 0 to 1 if Δ becomes sufficiently large. As in the basic setup, optimal prop- erty rights must be complete whenever v ≥ λm.
4.1.2. Market for Lemons
As noticed by Hasen and McAdams (1997, p. 372), “theft may avoid the ‘lem- ons problem.’” Here I formally study the conditions under which this is the case. For the sake of simplicity, I maintain that original owners have private informa- tion about v that is drawn from a uniform distribution with support [ , ] and correlated with the valuation of the potential buyers. In particular, Δ/2 potential buyers have a valuation of θv with θ > 2, 1 - Δ of them value x at v + μ with 0 < μ < (θ - 1)v, and the remainder have a valuation of v/θ. While θ measures the polarization of the potential buyer’s preferences, μ assumes the role of α in
Endogenous Property Rights 329
the basic setup. The price of x is exogenously fixed at pL > 2μ, and thus a poten- tial buyer infers that its expected value is pL/2 since an original owner sells only if v ≤ pL. While high-valuation potential buyers buy (expropriate) x if γ* ≥ (<) 2/θ, middle- and low-valuation buyers expropriate (do not consume) if γ* < (=) 1. Thus, society maximizes
( ) (1 ) (1 ) 4 2 2
(1 ) 2 2 2
for γ* ≥ 2/θ and
2
1 (1 ) (1 )
2 2 2 2
otherwise. As a consequence, the first-order condition of society’s problem is now independent of γ*, which can then jump from 0 to 1 provided that prefer- ence heterogeneity becomes sufficiently large.
4.2. Other Robustness Checks
4.2.1. Property Rules versus Liability Rules
After expropriation, a potential buyer has to return x with probability γ—that is, when the property rule is applied—and can keep it by paying damages d to the original owner otherwise—that is, when the liability rule is applied.13 I posit that there are frictions in the evaluation of d, which equals with probability ϕ and v otherwise.14 This assumption incorporates into the setup the legal system’s inability to estimate all the components of the original owner’s valuation (see Ka- plow and Shavell 1996, pp. 730–31) and/or its pro-defendant bias in response to the potential buyer’s guilt, shame, and indignation (Nicita and Rizzolli 2014). Hence, a potential buyer buys if R [ (1 ) /
ˆ ] ,v d v and she expropriates otherwise. This time, society maximizes
R
R
ˆ
ˆ
(1 )
2(1 )
(1 ) [(1 )(1 ) ] (1 ) [(1 )(1 ) ]
2(1 )
v
v
v d d
l l
v v d
l
for R
ˆ and
13 I use “liability rule” to refer to legal rules allowing transfers without consent but obliging the potential buyer to pay a weakly positive compensation to the original owner (Bar-Gill and Persico, forthcoming).
14 This setup also captures the case in which a state pays compensation after expropriation. A key extension to the model is to study the selection of property rights in anticipation of later bargaining (Segal and Whinston, forthcoming).
330 The Journal of LAW & ECONOMICS
m
(1 ) (1 ) [(1 )(1 ) ]
(1 ) [(1 )(1 ) ]
(1 )
v
v
v v d d
l l
v
otherwise. Social welfare under full expropriation decreases with γ for v < λm and decreases (possibly increases) with the strength of property rights for v ≥ λm and ϕ sufficiently small (large). For R
ˆ , society’s problem is convex in γ for ϕ sufficiently small and both concave and increasing in γ otherwise. Hence, γ* is 0 (possibly 1) for ϕ very small (large and v ≥ λm) and can jump from 0 to 1 as preference diversity becomes sufficiently large in the other cases. Intuitively, the frictionless liability rule with ϕ = 0 attains the second best by assuring that x is transferred to the potential buyer for certain at the end of λ ≥ v matches and stays with the original owner otherwise. As ϕ continues to rise, however, the pure liability rule misallocates more and more value after λ < v matches, and there- fore for v ≥ λm and both ϕ and Δ sufficiently large, society might prefer the pure property rule γ* = 1.
4.2.2. Heterogeneous (Homogeneous) Original Owners (Potential Buyers)
Assume that Δ/2 original owners value x at v > 0, Δ/2 of them value it at ,v v and the remainder have a valuation v uniformly distributed over [ , ]v v
with l v v and m ( )/2v v v . Potential buyers have the fixed valuation ( , )v v and can buy at v + α. Hence, only those matched with original own-
ers with ˆv v v buy, whereas the others expropriate. Notice that original owners sell when v is lower than λ - α, which in turn is higher than ˆ.v Society maximizes
ˆ
ˆ
(1 ) (1 ) (1 ) (1 ) ( )
2 2
v v
vv
v v dv dv
l l
for ˆv v and (1 - γ)λ + γvm otherwise. Since for γ* = 1 the potential buy- ers involved in λ - α < v ≤ λ matches are inefficiently excluded from trade, the analysis is similar to that of the basic setup. While γ* can jump from 0 to the positive solution, which in turn weakly rises with preference diversity, for Δ suf- ficiently large and vm < λ, γ* = 1 for vm ≥ λ.
4.2.3. Private Sales
A potential buyer has also the option of making a take-or-leave offer to her match and always consuming x if the offer is accepted. Thus, she transfers to her match γv and bears the cost α if P /
ˆ v and expropriates otherwise. Society maximizes
Endogenous Property Rights 331
P
P
ˆ
ˆ
( ) (1 ) (1 ) (1 ) (1 )
2 2
v v d d
l l
for P ˆ and W FE otherwise. In the former case, society’s problem is strictly
convex. As a consequence, γ* is either 1 or the corner solution 0 ( /( )v ) for v < λm (v ≥ λm), and the γ* = 1 case is more likely the larger preference diversity is under my hypotheses for α.
4.2.4. Costly Expropriation
Buying a stolen or embezzled good in good faith comes at a price, which is positive even if usually lower than that charged for goods with proper title (Dari-Mattiacci, Guerriero, and Huang 2016). Similarly, to occupy a plot or a house, the original owner’s protection effort needs to be circumvented. Here I use c > 0 to denote the socially wasteful cost borne by a potential buyer who expropriates. Potential buyers with C ( /
ˆ – )v c buy x, those with a valuation of
1 C(1
ˆ)c expropriate their match, and the remainder do not consume. Because the case in which
1 (1 )c is similar to the basic
setup, I assume that min{(1 ) , }.c Provided that 1
C(1 ) ˆ ,c 15 so-
ciety maximizes
C
C
ˆ
/(1 )ˆ
( ) (1 ) (1 ) (1 )
2
1
1 2
c
v c d d
l l
c v
l
for C ˆ and
/(1 )
(1 ) 1 (1 )
1 2
[(1 ) ] 2
c
v c c d v
l l
v c
otherwise. For C ˆ , the derivative of the social welfare function with respect
to γ is independent of Δ and positive. For C
ˆ instead, society’s problem is convex, and therefore γ* is either 0 or 1, with the latter more likely the larger Δ is for c not too small.
15 If 1
C(1 ) ˆ ,c then c ≥ (1 - γ)(v + α), all potential buyers buy, and property rights are
complete.
332 The Journal of LAW & ECONOMICS
4.2.5. Production
Other studies documenting that weak property rights can be optimal focus on endowment economies (Jordan 2006; Piccione and Rubinstein 2007; Bar-Gill and Persico, forthcoming). Here I prove that the mechanism highlighted in Sec- tion 3.2 survives when production is introduced into the framework. This time, the original owners decide between t1 and t2 whether to produce x at the cost κ < v. Production does not occur for ˆ , but it can be realized when the original owner’s expected utility
* ( )( )
(1 ) ( ˆ
1 ) 2
ˆ *
2
vv v v
l l
is weakly positive for ˆ and thus γ* > 0. Since such an expected utility rises with γ*, there is a , increasing with Δ for sufficiently large, such that x is produced only if * . Since production also creates value for the potential buyers, society always selects the maximum between and the solution to prob- lem (1).
4.2.6. Investment
The standard security argument in favor of property rights protection claims that expropriation induces a disincentive to commit effort or invest in technology (Besley and Ghatak 2010). To understand how this intuition affects the model’s message, I analyze an investment decision made by the original owners between t1 and t2 and shaping one of the three key ingredients of the model, in particular, either raising the original owner’s valuation to v(1 + ρ) with ( )/v v at the fixed cost ζ < v, increasing (decreasing) the probability that a potential buyer has valuation ( ) to Δ(1 + ρ) (Δ(1 - ρ)) at the fixed cost η < v,16 or allowing original owners to appropriate a share ρ of α at the variable cost ρ2/2 by, for instance, financing the potential buyer’s purchases. The analyses of the first two cases are similar to that in Section 4.2.5. Indeed, original owners invest only if I 0* (γ* > 0) and their expected utility
I
I I
I
(1 ) (1 )( ) (1 )
2
*(1 ) (1 )( ) * (1 )
2
ˆ
ˆ ,
v v
l
v v
l
where I I*[ (1 ) ]/ ˆ v and I* is defined by equation (2) with v(1 + ρ)
in place of v
16 Continuous versions of the first two exercises produce an algebra so tangled that it is unin- formative about the model’s robustness except when is large and v is small since then the basic analysis remains intact being dρ/dγ* small.
Endogenous Property Rights 333
ˆ(1 ) ( ) (1 )
2
(1 ) * ( ) * (1
ˆ ) ,
2
v v
l
v v
l
is weakly positive. Since this function rises with I* (γ*), there is a O ( B ), in- creasing with Δ for sufficiently large, such that the investment raising the original owner’s (potential buyer’s) valuation is made only if the protection of property rights is at least O ( B ). Society picks I* * instead of γ* when
OI * and social welfare is greater at I* with investment than it is at γ* with-
out investment, and O is preferred to γ* when OI* and social welfare is greater at
O with investment than it is at γ* without investment. Similarly, for
B* , society selects B instead of γ* when social welfare is greater at B with investment than it is at γ* without investment, and for B* society opts for γ*. Inducing investment thus weakly strengthens property rights protection, and investment always is made for large enough. Again, if v ≥ λm, then γ* = 1, and investment always materializes. When, finally, investment reduces socially wasteful transaction expenses, ρ* maximizes the concave
2
(1 )( )( ) ( )
2
(1 ) * (
ˆ
ˆ ) *
2 2
v v
l
v v
l
function for ˆ and γ*v - ρ2/2 otherwise. Hence, the optimal investment level is
ˆ( ) (1 ) 1
2 l
and rises with γ* for ˆ , and it is 0 otherwise. Insecure property rights de- press investment, and society considers dynamic inconsistency at t0. The first- order condition of society’s problem for ˆ equals that in Section 3.2 except for a positive investment inducement term. For v < λm then, γ* can jump from 0 to a positive γ* higher than that prevailing without investment and rising with Δ for Δ large enough. Once again, v ≥ λm implies that γ* = 1.
5. Evidence
The model implies that a rise in preference diversity leads to stronger property rights since it reduces the mass of middle-valuation potential buyers inefficiently excluded from trade because of transaction costs and at the same time increases the mass of low-valuation potential buyers who inefficiently expropriate the orig- inal owners. This intuition is true whether or not constitution writers choose un- der a veil of ignorance and implies the following prediction:
334 The Journal of LAW & ECONOMICS
Prediction 1. The strength of property rights will increase with preference het-
erogeneity.
5.1. Identifying Causal Relationships
The positive correlations between horizontal property rights and contem-
porary genetic diversity documented in Section 2 are consistent with the main
prediction of the model. Nevertheless, they could also be explained by omitted
variables. Empirical works building on the primordial theory of social evolution
document that contemporary preference heterogeneity was produced mainly by
the prehistoric exodus of humans out of Africa and the consecutive agricultural
revolution, but those in line with the constructive theory also emphasize the
European colonizers’ attempt to divide and rule the natives (Ashraf and Galor
2013). Thus, it is unlikely that the results are driven by reverse causation, but it is
possible that they spuriously reflect omitted variables, such as colonial rule, that
are correlated with preference diversity and institutions, like an inclusive politi-
cal process and state capacity (Acemoglu and Johnson 2005), that are conducive
to stronger property rights.
I pursue four strategies to assess if the correlations documented so far are
causal. First, I control for colonial rule and other determinants of property rights.
Second, I focus on the jurisdictions in which the population structure has not
been heavily modified by colonization and thus the majority of contemporary
inhabitants belong to the native population. Third, I use selection on observables
to assess the likelihood that the estimates are driven by unobservables. Finally, I
turn to a 2SLS estimator using the prehistoric migratory distance from East Af-
rica as the excluded instrument for contemporary genetic diversity.
In the online appendix, I provide a number of further robustness checks. First,
allowing for clustering by continent to take into account the within-continent
correlation in the error term produced by legal harmonization at the level of
political and commercial unions implies extremely similar estimates. Second,
since preference heterogeneity might be relevant only when market networks are
dense, and as a result potential buyers with very different valuations are matched
to original owners, I also estimate each specification using weighted least squares
regression with weights proportional to population density. The conclusions of
this exercise are consistent with the model’s prediction. Finally, I obtain similar
results when I use a probit model to estimate the specifications with dependent
variable Government- Takings.
5.1.1. Controlling for Observables
The key variables potentially omitted from the analysis in Section 2 are those shaping property rights through channels other than the trade-off between ineffi- cient exclusion from trade and inefficient expropriation and associated with pref- erence diversity. They are the determinants of the Europeans’ settlement strategy, development level, inclusiveness of political institutions, drivers of state capac-
Endogenous Property Rights 335
ity, internal conflicts, and geographic features shaping preference heterogeneity and the economy. I illustrate these in turn.
First, Acemoglu, Johnson, and Robinson (2001, 2002) contend that European colonizers invested in institutions of private property and preserved ethnic di- versity where the disease environment allowed for more prosperous settlement and in less populated regions where coercion of the natives was a suboptimal option. Thus, discarding these two factors as irrelevant could lead me to falsely attribute to Genetic-Diversity-A the role of colonial rule. Hence, I consider the historical prevalence of seven disease-causing pathogens, Pathogen-Load; the natural logarithm of the population density in 1500, Population-D-1500; and the duration of the colonization experience, Colonization-Time.
Second, preference heterogeneity could affect property rights through its im- pact on development and in particular by modulating the modernization effect (Ashraf and Galor 2013). Hence, I consider the natural logarithm of the 2010 population, Population, and of the 2010 real GDP per capita relative to the United States, Income. Controlling for these two observables also takes into ac- count the differences between developing and developed jurisdictions.
Third, Aghion, Alesina, and Trebbi (2004) propose the idea that in fragmented societies, while it would be optimal to guarantee a voice to everybody, power- ful minorities impose extractive institutions, weakening private rights. To assess the importance of this mechanism, which is, however, inconsistent with the esti- mates discussed so far, I use the constraints on the executive authority score from the POLITY IV data set, Democracy, and the time since the agricultural tran- sition, Neolithic.17 Olsson and Paik (2016) relate the latter to more patriarchal present-day values that, in turn, delay democratization.
Fourth, a broad literature claims that low levels of nonproduced output, which magnifies the disincentive effect of expropriation, the historical incidence of ex- ternal wars, and German and Scandinavian legal origins are remarkably stable predictors of a jurisdiction’s capacity to control expropriation and protect pri- vate rights (Besley and Persson 2009). Thus, I evaluate the role of the following three controls: the 2010 crude oil proved reserves per capita, Reserves; the share of years between 1823 and 2007 in which the jurisdiction was involved in an ex- ternal conflict, Conflict-E; and a continuous measure of the decentralization of the jurisdiction’s legal tradition in 2000 ranging between 0 and 1 and assuming higher values the nearer it is to a perfectly decentralized common-law system, Common-Law.18
17 The Democracy score ranges between 1 and 7 and is higher when the holder of the executive power is accountable to the citizens and/or the government is constrained by checks and balances and/or the rule of law.
18 Common-Law represents the normalized first principal component extracted from a dummy variable for case law, a binary that turns on if the judgment may be based on both law and equity grounds, a dummy variable equal to one if only new evidence or issues of law can be reviewed or if there is no appeal, an indicator variable equal to one if judges cannot freely request or hear evidence that has not been introduced by the parties and cannot refuse to collect or admit requested evidence —.5 if only one of these two bans is imposed and zero otherwise, and a binary variable that turns on if the evidence is mostly submitted at oral hearings before a judge (Guerriero 2016). The results will be similar if I use the La Porta et al.’s (1999) English common-law dummy variable, instead.
336 The Journal of LAW & ECONOMICS
Fifth, preference and in particular ethnic diversity might adversely affect prop- erty rights protection by fostering intergroup conflict and thus expropriation (Cervellati, Fortunato, and Sunde 2008). Although at odds with the preliminary conclusions of Section 2, I consider this possibility by examining the share of years between 1816 and 2007 in which the jurisdiction experienced an internal mili- tary conflict, Conflict-I. A mechanism that is instead consistent with the evidence summarized in Table 2 is the one through which contemporary genetic diversity fosters property rights protection by increasing the likelihood of distrust (Ashraf and Galor 2013), which in turn requires centralized enforcement as a substitute for malfunctioning relational contracts (Aghion et al. 2010).19 Crucially, a culture of morality and the efficiency of public enforcement also alter the functioning of the intermediation technology and thus the balance between protecting property and enhancing reliance on contracts (Dari-Mattiacci and Guerriero 2015). Esti- mates reported in the online appendix hint at the contemporaneous relevance of these alternative mechanisms and the model’s testable prediction.20 Considering all of the controls introduced so far also factors in differences in transaction costs driven by conflicts, the inclusiveness of institutions, and culture (La Porta et al. 1999). Since transaction costs might be correlated with preference heterogeneity, this aspect of the identification strategy is particularly important.
Finally, Michalopoulos (2012) documents that geographic variability, as re- flected by variation in the quality of the land, elevation, and climate, contrib- uted to ethnolinguistic fractionalization, whereas Boranbay and Guerriero (2015) show that the same features have persistently determined culture and political institutions in Europe. Accordingly, I take into account the standard deviation of the quality of the land, Land-Quality-SD; its ruggedness, Ruggedness; and the normalized absolute latitude, Latitude. In the online appendix, I also experiment with the annual mean temperature and precipitation averaged over 1961–1990, and I obtain similar results.
Columns 2–8 in Table 2 report the estimates from the specifications consider- ing first stepwise and then together the extra controls. The key observation is that the coefficients on these regressors are rarely statistically significant, whereas that on Genetic-Diversity-A is positive and significant at 10 percent or better and has a magnitude that is almost equal whether none or all of the controls are included. The scatterplots in Figure 3 display the estimates of Genetic- Diversity-A from the specifications controlling for all observables and reveal that they are not driven by abnormal observations.21
19 I thank James Dana and an anonymous reader for drawing my attention to this point. 20 While I proxy for culture using the first principal component extracted from the level of gen-
eralized trust and the importance of respect self-reported by respondents to the World Value Sur- veys and European Value Studies before 2008, I measure the efficiency of public enforcement with the first principal component extracted from the number of police officers and that of professional judges per 100,000 inhabitants averaged between 1973 and 2009.
21 Accordingly, I document in the online appendix that the evidence remains essentially the same when I exclude the outliers identified with Cook’s distance (Cook 1977).
Endogenous Property Rights 337
5.1.2. Focusing on Native Populations
The estimates in Table 2 do not exclude that colonial rule is driving contempo- rary genetic diversity and property rights via a channel other than the Europeans’ settlement strategy. To determine if this is the case, I repeat the OLS analysis dis- cussed so far, employing only jurisdictions in which more than 50 percent of the current population was indigenous as of 1500 according to the data in Putterman and Weil (2010). Table 3 clearly shows that the basic results do not rest on the inclusion of jurisdictions where genetic pools have been heavily modified by Eu- ropean colonization.
5.1.3. Using Selection on Observables to Assess the Bias from Unobservables
Despite my attempts to control for relevant observable factors, my OLS esti- mates may still be biased by unobservables. To evaluate this issue, I calculate the index proposed by Altonji, Elder, and Taber (2005) to measure how much stron- ger selection on unobservables, relative to selection on observables, must be to explain away the full estimated effect.22 To see how the index is calculated, con- sider a regression with a restricted set of control variables and one with a full set of controls. Next denote the estimate of the coefficient on Genetic-Diversity-A from the first regression βR, where R is restricted, and that from the second re- gression βF, where F is full. Then the index is the absolute value of βF/(βR - βF). The intuition behind the formula is as follows. The lower the absolute value of (βR - βF) is, the less the estimate of the relevant coefficient is affected by selection on observables and the stronger selection on unobservables needs to be to explain away the entire effect. Moreover, the higher the absolute value of βF is, the greater is the effect that needs to be explained away by selection on unobservables and thus the higher the index is.
In Table 4, I consider the specification without additional controls in column 1 of Table 2 as the restricted regression and those in columns 2–8 of Table 2 as the full regressions. The ratios calculated when the dependent variable is respectively AP-Movable, AP-Immovable, and Government-Takings are reported in columns (1)–(3). None of them is less than one, and the median and average ratios ob- tained from the full regressions including all observables in column (8) of Table 2 are 5.4 and 12.5. Therefore, to attribute the entire OLS estimate of the coefficient on Genetic- Diversity-A to selection effects, selection on unobservables would have to be on average more than twelve times greater than selection on all observ- ables. Given the high fit of the OLS regressions including all observables in Table 2, I conclude that such a circumstance is quite unlikely.
22 I use the version developed by Bellows and Miguel (2009) for possibly endogenous continuous variables.
338
T a
b le
2
E n
d o
g e
n o
u s
P ro
p e
rt y
R ig
h ts
: O
rd in
a r
y L
e a
st S
q u
a re
s E
st im
a te
s
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
A P
-M o v
a b
le (
N =
8 1
):
G
en et
ic -D
iv er
si ty
-A 1
1 8
.2 0
2 **
(3 0
.6 1
2 )
1 5
3 .1
4 9
** (3
4 .6
7 2
) 1
2 8
.0 5
0 **
(2 6
.0 0
2 )
1 3
2 .5
8 3
** (3
4 .7
1 1
) 1
0 6
.3 3
2 **
(3 2
.7 0
5 )
1 3
0 .2
3 0
** (3
1 .3
8 8
) 1
4 4
.6 9
2 **
(3 0
.8 1
6 )
1 1
4 .0
6 3
** (4
0 .8
6 5
)
P
a th
o g
en -L
o a d
1 .5
3 0
(2 .3
6 6
)
P
o p
u la
ti o
n -D
-1 5
0 0
- 1
.5 5
9 +
(. 8
3 6
)
C
o lo
n iz
a ti
o n
-T im
e .4
4 6
(. 6
4 3
)
P
o p
u la
ti o
n 1
.1 4
3 (.
8 8
2 )
In
co m
e -
1 .6
4 4
(1 .1
2 0
)
N
eo li
th ic
- 1
.3 9
1 *
(. 6
7 6
)
D
em o
cr a cy
.9 6
7 (.
7 8
0 )
R
es er
v es
- 3
.5 E -
1 1
(1 .4
E -
9 )
C
o n
fl ic
t- E
- 1
5 .8
3 7
(3 8
.5 3
9 )
C
o m
m o
n -L
a w
1 8
.4 0
8 **
(4 .3
0 9
)
C
o n
fl ic
t- I
2 0
.0 5
3 (1
9 .9
4 1
)
L
a n
d -Q
u a li
ty -S
D 6
.8 8
3 (1
4 .8
9 0
)
R
u g
g ed
n es
s -
1 3
.5 1
7 (8
.8 1
9 )
L
a ti
tu d
e -
1 7
.2 8
4 *
(7 .1
3 7
)
E
x tr
a c
o n
tr o
ls (
p -v
a lu
e) .0
0
R
2 .0
9 .1
6 .1
5 .1
8 .2
7 .1
0 .1
7 .4
9
339
A P
-I m
m o v
a b
le (
N =
6 3
):
G
en et
ic -D
iv er
si ty
-A 7
4 .1
3 1
* (3
4 .7
2 2
) 8
6 .3
3 4
* (3
3 .5
9 2
) 8
7 .8
6 5
* (3
4 .9
5 7
) 9
6 .3
7 6
** (3
3 .7
2 2
) 7
7 .7
5 2
* (3
5 .9
3 8
) 9
0 .7
6 5
** (3
3 .1
9 2
) 9
4 .3
9 6
* (3
6 .1
6 7
) 9
1 .0
4 7
+
(4 9
.1 2
3 )+
P
a th
o g
en -L
o a d
3 .8
6 1
* (1
.5 4
0 )
P
o p
u la
ti o
n -D
-1 5
0 0
.0 6
8 (.
5 6
4 )
C
o lo
n iz
a ti
o n
-T im
e -
.2 4
7 (.
4 2
9 )
P
o p
u la
ti o
n .7
2 3
(. 6
1 0
)
In
co m
e -
1 .5
8 9
* (.
6 8
6 )
N
eo li
th ic
- .2
0 7
(. 3
6 5
)
D
em o
cr a cy
- 1
.0 5
2 (.
6 6
9 )
R
es er
v es
- 5
E -
9 **
(9 .5
E -
1 0
)
C
o n
fl ic
t- E
- 1
3 .5
0 9
(2 9
.0 7
9 )
C
o m
m o
n -L
a w
- .8
4 2
(3 .7
4 3
)
C
o n
fl ic
t- I
2 4
.5 0
0 *
(1 1
.1 2
0 )
L
a n
d -Q
u a li
ty -S
D -
6 .7
0 5
(1 1
.6 6
7 )
R
u g
g ed
n es
s 2
.1 0
7 (8
.2 1
4 )
L
a ti
tu d
e -
5 .2
5 6
(5 .4
0 3
)
E
x tr
a c
o n
tr o
ls (
p -v
a lu
e) .0
0
R
2 .0
6 .1
3 .1
3 .1
0 .0
9 .1
0 .0
8 .2
5
340
G o v
er n
m en
t- T
a k
in g
s (N
= 6
2 ):
G
en et
ic -D
iv er
si ty
-A 3
.9 8
5 +
(2 .1
1 5
) 4
.4 3
6 +
(2 .4
1 9
) 3
.5 0
6 (2
.2 0
9 )
5 .2
3 4
* (2
.2 7
9 )
3 .7
1 6
+
(2 .0
3 7
) 2
.9 5
8 (2
.3 2
1 )
4 .1
1 5
+
(2 .2
3 0
) 5
.0 9
5 *
(2 .3
8 8
)
P a th
o g
en -L
o a d
- .1
1 4
(. 0
9 4
)
P o
p u
la ti
o n
-D -1
5 0
0 -
.0 6
8 *
(. 0
2 9
)
C o
lo n
iz a ti
o n
-T im
e .0
0 6
(. 0
3 5
)
P o
p u
la ti
o n
- .0
3 5
(. 0
4 2
)
In co
m e
.0 3
4 (.
0 5
5 )
N
eo li
th ic
- .0
3 7
(. 0
2 5
)
D em
o cr
a cy
- .0
2 0
(. 0
3 5
)
R es
er v
es 4
.5 E -
1 1
( 7
E -
1 1
)
C o
n fl
ic t-
E -
2 .3
5 8
(1 .9
4 8
)
C o
m m
o n
-L a w
.0 7
7 (.
2 0
8 )
C
o n
fl ic
t- I
- 1
.4 6
9 *
(. 6
1 5
)
L a n
d -Q
u a li
ty -S
D -
.9 6
3 (.
5 8
4 )
R
u g
g ed
n es
s -
.3 5
3 (.
4 9
0 )
L
a ti
tu d
e -
.1 1
5 (.
3 1
0 )
E
x tr
a c
o n
tr o
ls (
p -v
a lu
e) .1
1
R 2
.0 6
.1 3
.0 8
.0 9
.0 9
.1 1
.1 1
.2 4
S o
u rc
e.
S ee
T a b
le 1
f o
r ea
ch v
a ri
a b
le d
efi n
it io
n a
n d
s o
u rc
es .
N o
te .
R o
b u
st s
ta n
d a rd
e rr
o rs
i n
p a re
n th
es es
. A
ll s
p ec
ifi ca
ti o
n s
in cl
u d
e a c
o n
st a n
t te
rm ,
a n
d t
h o
se i
n c
o lu
m n
8 a
ls o
i n
cl u
d e
P a th
o g
en -L
o a d
, P
o p
u la
ti o
n -D
-1 5
0 0
, C
o lo
n iz
a ti
o n
-T im
e, P
o p
u la
ti o
n , In
co m
e, N
eo li
th ic
, D
em o
cr a cy
, R
es er
v es
, C
o n
fl ic
t- E
, C
o m
m o
n -L
a w
, C
o n
fl ic
t- I,
L a n
d -Q
u a li
ty -S
D , R
u g
g ed
n es
s, a
n d
L a ti
tu d
e. + p
< .
1 0
. *
p <
. 0
5 .
** p
< .
0 1
.
T a
b le
2 (
C o
n ti
n u
ed )
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
341
Figure 3. Endogenous property rights: ordinary least squares coefficients
342 The Journal of LAW & ECONOMICS
5.1.4. Two-Stage Least Squares Estimates
My final strategy is to use instrumental variables. This requires an instrument that is correlated with contemporary genetic diversity but uncorrelated with any other characteristic of the jurisdiction affecting horizontal property rights. I rely on the prehistoric migratory distance from the cradle of humankind, identified by a large population genetics literature as Addis Ababa, Migratory- Distance (Ashraf and Galor 2013). This variable is calculated as a great-circle distance from Addis Ababa to the centroid of each jurisdiction along a path touching one or more obligatory intermediate waypoints that synthesize paleontological evi- dence of prehistoric migration patterns. Of course, it is quite difficult to envi- sion any systematic relationship between prehistoric migration patterns driven by prehistoric natural disasters, such as volcanic eruptions and glaciations, and present-day property rights once I control for the key determinants of the Euro- pean colonizers’ settlement strategy, development level, inclusiveness of political institutions, external and internal conflicts, other drivers of state capacity such as legal origins, and geographic variability.
Table 5 reports in panels A–C the 2SLS estimates of the specifications with de- pendent variable respectively AP-Movable, AP-Immovable, and Government- Takings. Starting from the first stages, Migratory-Distance has always a negative
and significant at 1 percent coefficient and explains alone at least 59 percent of the variability in Genetic-Diversity-A (not shown). Accordingly, I can always reject that the estimated equation is underidentified at 1 percent. Turning to the second stages, the coefficient on Genetic-Diversity-A is always positive and strongly sig- nificant when I control for all observables. Crucially, the estimates of the specifi- cations including all observables in Table 5 and their OLS counterparts in Table 2 have almost the same magnitude. This evidence entails that the potential en-
Table 3
Endogenous Property Rights: Ordinary Least Squares Estimates Obtained Using Native Population Samples
AP-Movable (N = 64)
AP-Immovable (N = 47)
Government-Takings (N = 46)
(1) (2) (3) (4) (5) (6)
Genetic-Diversity-A 98.287** 106.953+ 53.771 176.904** 2.197 8.106*
(35.849) (61.427) (39.095) (54.720) (2.494) (3.676)
Extra controls (p-value) .00 .00 .00
R2 .06 .48 .03 .37 .02 .33
Source. See Table 1 for each variable definition and sources. Note. Robust standard errors in parentheses. All specifications include a constant term, and those in columns 2, 4, and 6 also include Pathogen-Load, Population-D-1500, Colonization-Time, Population, Income, Neolithic, Democracy, Reserves, Conflict-E, Common-Law, Conflict-I, Land- Quality- SD, Ruggedness, and Latitude.
+ p < .10. * p < .05. ** p < .01.
Endogenous Property Rights 343
dogeneity between contemporary genetic diversity and property rights need not be a source of concern here. Consistent with this view, I cannot reject that Ge- netic-Diversity-A is exogenous conditional on all observables at a level nowhere lower than 5 percent.23 Overall, these patterns provide support for the inferences made above and confirm the validity of the model’s testable prediction.
5.1.5. Alternative Measures of Preference Heterogeneity
I conclude my test of the model’s prediction by documenting that the basic results hold when I switch to different measures of preference heterogeneity. I consider in Table 6 either the aggregate expected heterozygosity in precolonial indigenous settlements as predicted by the prehistoric migratory distance from East Africa only, Genetic-Diversity, or one of the three proxies for cultural diver- sity most widely used by the extant literature on the impact of heterogeneity on economic outcomes (Alesina et al. 2003; Michalopoulos 2012): the natural log- arithm of the number of languages, Language; the ethnic fractionalization, Eth- nic; and the religious fractionalization, Religion. The latter two variables gauge the probability that two randomly selected individuals in the population differ in their ethnicity and religion.
My estimates confirm that preference heterogeneity, whether produced by dif- ferent genetic pools, diverse ethnic roots, different communication tools, or di-
23 Conditional on all observables, I cannot reject that the overidentifying restrictions hold at a level nowhere lower than 17 percent when I use as a second excluded instrument the time from the first uninterrupted human settlement in 100,000s of years (results available on request). Ahlerup and Olsson (2012) document that this variable is strongly correlated with ethnic diversity since during the Neolithic revolution new ethnic groups progressively emerged among peripheral popu- lations as a response to an insufficient supply of public goods.
Table 4
Using Selection on Observables to Assess the Bias from Unobservables
AP- Movable (N = 81)
AP- Immovable (N = 63)
Government- Takings
(N = 62)
The full set of covariates includes a constant term and
Pathogen-Load, Population-D-1500, and Colonization-Time 4.382 7.075 9.836
Population and Income 13.003 6.398 7.319
Neolithic and Democracy 9.219 4.332 4.191
Reserves, Conflict-E, and Common-Law 8.958 21.473 13.814
Conflict-I 10.827 5.457 2.880
Land-Quality-SD, Ruggedness, and Latitude 5.462 4.658 31.654
All extra controls 27.558 5.382 4.590
Source. See Table 1 for each variable definition and sources. Note. Each cell reports an index constructed as explained in Section 5.1.3 and based on the coef- ficients attached to Genetic- Diversity-A and obtained from two regressions. In one, the covariates include only a constant term and Genetic-Diversity-A. In the other, the full set of covariates gathers these two variables and the controls indicated by the row labels.
T a
b le
5
E n
d o
g e
n o
u s
P ro
p e
rt y
R ig
h ts
: T
w o
-S ta
g e
L e
a st
S q
u a
re s
E st
im a
te s
(1 )
(2 )
(3 )
(4 )
(5 )
(6 )
(7 )
(8 )
A .
A P
-M o v
a b
le (
N =
8 1
):
G
en et
ic -D
iv er
si ty
-A 1
4 1
.0 3
4 **
2 6
3 .8
0 6
** 1
4 3
.8 0
9 **
1 8
5 .9
0 1
** 1
1 0
.7 9
9 **
1 5
2 .0
2 0
** 1
9 3
.2 5
9 **
2 0
2 .1
7 2
**
(4 1
.6 9
3 )
(5 5
.8 7
5 )
(3 8
.8 7
1 )
(4 5
.9 3
1 )
(3 9
.3 2
8 )
(4 4
.5 5
8 )
(5 1
.2 5
7 )
(6 1
.9 9
5 )
F
ir st
s ta
g e:
M ig
ra to
ry -D
is ta
n ce
- .0
0 3
1 **
- .0
0 3
7 **
- .0
0 3
1 **
- .0
0 3
3 **
- .0
0 3
1 **
- .0
0 3
0 **
- .0
0 3
0 **
- .0
0 3
6 **
(. 0
0 0
4 )
(. 0
0 0
4 )
(. 0
0 0
5 )
(. 0
0 0
4 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
6 )
E
n d
o g
en ei
ty (
p -v
a lu
e) .4
3 .0
1 .5
9 .1
0 .8
7 .4
8 .2
4 .0
5
B .
A P
-I m
m o v
a b
le (
N =
6 3
):
G
en et
ic -D
iv er
si ty
-A 9
6 .5
5 5
* 1
2 5
.2 9
5 **
1 0
6 .2
5 7
* 1
4 8
.2 6
5 **
1 0
1 .5
1 8
* 1
1 0
.3 9
2 *
1 4
0 .4
0 1
** 1
3 5
.6 2
2 *
(4 5
.4 7
4 )
(4 4
.8 3
7 )
(4 6
.6 5
0 )
(4 4
.2 2
3 )
(4 5
.3 6
3 )
(4 7
.1 5
4 )
(5 1
.5 2
2 )
(6 7
.4 6
2 )
F
ir st
s ta
g e:
M ig
ra to
ry -D
is ta
n ce
- .0
0 2
9 **
- .0
0 3
5 **
- .0
0 2
8 **
- .0
0 2
9 **
- .0
0 2
9 **
- .0
0 2
8 **
- .0
0 2
7 **
- .0
0 3
3 **
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
6 )
(. 0
0 0
7 )
E
n d
o g
en ei
ty (
p -v
a lu
e) .4
0 .2
0 .5
2 .1
1 .3
5 .5
0 .1
8 .3
2
C .
G o
v er
n m
en t-
T a k
in g
s (N
= 6
2 ):
G
en et
ic -D
iv er
si ty
-A 1
.9 7
4 4
.4 4
0 +
1 .6
0 5
3 .8
2 7
1 .7
6 4
1 .1
1 7
2 .0
5 4
8 .6
6 4
**
(2 .5
5 3
) (2
.7 5
0 )
(2 .6
2 7
) (2
.8 9
1 )
(2 .5
1 8
) (2
.6 9
6 )
(2 .9
7 8
) (2
.8 1
1 )
F
ir st
s ta
g e:
M ig
ra to
ry -D
is ta
n ce
- .0
0 2
9 **
- .0
0 3
5 **
- .0
0 2
8 **
- .0
0 2
9 **
- .0
0 2
9 **
- .0
0 2
8 **
- .0
0 2
7 **
- .0
0 3
4 **
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
5 )
(. 0
0 0
6 )
(. 0
0 0
7 )
E
n d
o g
en ei
ty (
p -v
a lu
e) .1
5 .9
9 .1
9 .4
1 .1
6 .2
1 .2
9 .0
8
S o
u rc
e.
S ee
T a b
le 1
f o
r ea
ch v
a ri
a b
le d
efi n
it io
n a
n d
s o
u rc
es .
N o
te .
R o
b u
st s
ta n
d a rd
e rr
o rs
i n
p a re
n th
es es
. A
ll s
p ec
ifi ca
ti o
n s
in cl
u d
e a c
o n
st a n
t te
rm .
Th e
sp ec
ifi ca
ti o
n s
in c
o lu
m n
s 2
– 8
a ls
o i
n cl
u d
e (1
) P
a th
o g
en - L
o a d
, P
o p
u la
ti o
n -D
-1 5
0 0
, a n
d
C o
lo n
iz a ti
o n
-T im
e;
(2 )
P o
p u
la ti
o n
a n
d
In co
m e;
(3
) N
eo li
th ic
a n
d
D em
o cr
a cy
; (4
) R
es er
v es
, C
o n
fl ic
t- E
, a n
d
C o
m m
o n
-L a w
; (5
) C
o n
fl ic
t- I;
( 6
) L
a n
d -Q
u a li
ty -S
D ,
R u
g g
ed n
es s,
a n
d L
a ti
tu d
e; a
n d
( 7
) a ll
e x
tr a c
o n
tr o
ls .
Th e
n u
ll h
y p
o th
es is
o f
th e
K le
ib er
g en
-P a a p
t es
t is
t h
a t
M ig
ra to
ry -
D is
ta n
ce i
s u
n co
rr el
a te
d w
it h
G en
et ic
-D iv
er si
ty -A
, a n
d t
h e
re la
ti v
e p
-v a lu
e is
. 0
0 f
o r
a ll
s p
ec ifi
ca ti
o n
s. Th
e n
u ll
h y
p o
th es
is o
f th
e en
d o
g en
ei ty
t es
t is
t h
a t
G en
et ic
- D
iv er
si ty
-A c
a n
b e
tr ea
te d
a s
ex o
g en
o u
s. + p
< .
1 0
. *
p <
. 0
5 .
** p
< .
0 1
.
Endogenous Property Rights 345
verse religious values, always reduces the need for welfare- enhancing noncon- sensual transfers. These results are even more remarkable given that religious fractionalization captures a source of variation in preference diversity different from that synthesized by the other metrics as revealed by its low—that is, no- where higher than .31—correlation with them.
Table 6
Endogenous Property Rights: Alternative Measures of Preference Heterogeneity
AP-Movable (N = 81)
AP-Immovable (N = 63)
Government- Takings
(N = 62)
OLS (1)
2SLS (2)
OLS (3)
2SLS (4)
OLS (5)
2SLS (6)
Genetic-Diversity 90.073** 98.718** 70.883* 61.005* 3.496* 4.033**
(31.525) (27.732) (35.685) (31.647) (1.735) (1.438)
R2 .51 .26 .24
Underidentification (p-value) .00 .00 .00
Endogeneity (p-value) .23 .03 .04
Language 1.208 9.611* .492 5.526 .091 .393*
(.992) (3.881) (1.101) (3.483) (.066) (.201)
R2 .46 .20 .22
Underidentification (p-value) .00 .01 .01
Endogeneity (p-value) .01 .08 .06
Ethnic 4.191 56.983* .846 75.849 .143 4.491
(5.620) (22.815) (6.531) (77.362) (.393) (3.590)
R2 .46 .20 .20
Underidentification (p-value) .02 .21 .18
Endogeneity (p-value) .00 .08 .02
Religion 11.027+ 55.348** -4.809 35.425 .379 2.625*
(5.939) (19.991) (6.936) (24.400) (.313) (1.365)
R2 .49 .21 .21
Underidentification (p-value) .01 .03 .06
Endogeneity (p-value) .01 .03 .03
Source. See Table 1 for each variable definition and sources. Note. Robust standard errors in parentheses. All specifications include a constant term, Pathogen- Load, Population-D-1500, Colonization-Time, Population, Income, Neolithic, Democracy, Reserves, Conflict-E, Common-Law, Conflict-I, Land-Quality-SD, Ruggedness, and Latitude. The null hypoth- esis of the Kleibergen-Paap test is that Migratory-Distance is uncorrelated with the endogenous re- gressor. The null hypothesis of the endogeneity test is that the endogenous regressor can be treated as exogenous.
+ p < .10. * p < .05. ** p < .01.
346 The Journal of LAW & ECONOMICS
5.2. Interpreting the Empirical Results
Overall, it is fair to view the empirical results as consistent with, if not prov- ing, causality running from preference diversity to horizontal property rights. A key issue, however, remains: how do private-taking-based measures of property rights relate to the subjective proxies for the risk of predation by the state and elites that are currently used in the literature?
To answer this question, I correlate the measures of horizontal property rights used above with the two proxies for vertical property rights most widely used in the literature (Acemoglu and Johnson 2005): Private-Property, which is a five-category index coded by the Heritage Foundation with higher values indi- cating more perceived protection of private property from government expro- priation in 1997, and Av-Expropriation, which is a 10-category index produced by Political Risk Services with higher values indicating lower perceived risk of expropriation of private foreign investment by the government and powerful elites between 1985 and 1995. A glance at Table 7 immediately reveals that the two predation-based proxies for property rights have a small and possibly neg- ative correlation with the private-taking-based measures of property rights. This pattern implies that one should be careful about drawing any general conclusion about the economic role of property rights protection from correlations between predation-based proxies and economic outcomes.
6. Conclusions
This paper develops a theory of endogenous property rights that also explains how much different societies protect property depending on their preference di- versity. I study a simple yet general society which is equally split into original owners, who have the same valuation for the only good in the economy, and po- tential buyers, who are instead endowed with heterogeneous valuations. The for- mer are randomly matched to the latter by an intermediation technology that allows each potential buyer to either obtain the good via consensual transfer by paying the original owner’s valuation and bearing a socially wasteful transac- tion cost or to expropriate it at no cost. I define property rights as the probability that an expropriated good is returned to its original owner. When property is
Table 7
Correlations among Measures of Property Rights
Private- Property
Av- Expropriation
AP- Movable
AP- Immovable
Av-Expropriation .70
AP-Movable .09 .14
AP-Immovable -.17 -.19 .21
Government-Takings -.04 -.11 -.07 .15
Source. See Table 1 for each variable definition and sources. Note. N = 64.
Endogenous Property Rights 347
fully protected, some potential buyers with valuation higher than that of origi- nal owners are inefficiently excluded from trade because of the transaction cost. When protection of property is weak, low-valuation potential buyers inefficiently expropriate original owners. A rise in the heterogeneity of the potential buyer’s valuations makes inefficient expropriation by low-valuation potential buyers weigh on social welfare more heavily than inefficient exclusion from trade and thus induces stronger property rights. This prediction survives if the economy becomes political and thus someone is excluded from the institutional design, when the transaction cost is endogenously determined by either market power or asymmetric information, when potential buyers need to pay damages to keep a successfully expropriated good, when original owners have heterogeneous valua- tions, when individuals can bargain over the price, when expropriation is costly, and when production and investment activities are considered.
To evaluate the model’s prediction, I exploit a novel data set on the rules on the acquisition of ownership through adverse possession and on the use of gov- ernment takings to transfer real property from a private party to another pre- vailing in 126 jurisdictions. Building on population genetics and recent studies of endogenous diversity, I proxy preference heterogeneity with contemporary genetic diversity as predicted by the prehistoric migratory distance from East Af- rica and post-1500 population flows. My OLS estimates suggest that protection of the original owner’s property rights is the strongest where contemporary ge- netic diversity is the largest. To determine if this relationship is indeed causal, I pursue several strategies. First, I document that the gist of the empirical exercise is robust to considering the key determinants of the colonizers’ settlement strat- egy, development level, inclusiveness of political institutions, external and inter- nal conflicts, other drivers of state capacity such as legal origins, and geographic features that affect preference heterogeneity and the structure of the economy. Second, I show that the estimates are similar when I focus on jurisdictions in which the population structure has not been heavily modified by colonization and thus the majority of contemporary inhabitants can trace their ancestry back to 1500 to the same area. Third, I build on Altonji, Elder, and Taber (2005) and I calculate that selection on unobservables would have to be on average more than twelve times greater than selection on all observables for the positive effect of contemporary genetic diversity on property rights to be spurious. Given the high fit of the regressions, this is unlikely. Finally, I exploit the plausible exogeneity of the prehistoric migratory distance from East Africa and I use it as an excluded in- strument for contemporary genetic diversity. The 2SLS estimates are qualitatively identical to the OLS estimates, and therefore I cannot reject that contemporary genetic diversity is exogenous at a level nowhere lower than 5 percent, condi- tional on all observable factors.
I close by highlighting two key avenues for future research. First, the tendency of property rights toward optimality does not imply that the existing legal vari- ation in the rules on private takings is irrelevant and thus does not warrant re- forms. On the contrary, the model reveals that special interests can distort the
348 The Journal of LAW & ECONOMICS
design of property rights away from optimality when the political process is less than perfect. Thus, a key issue to investigate is the impact of regulation of pri- vate takings on outcomes. Second, since the private-taking-based measures of property rights correlate weakly and possibly negatively with the most used predation-based proxies for property rights, further research on the relationships among horizontal property rights, vertical property rights, and the economy is needed.
Appendix A
Countries in the Full Sample
Albania, Angola, Argentina, Armenia, Australia, Austria, Azerbaijan, Bahrain, Bangladesh, Belarus, Belgium, Belize, Bolivia, Bosnia and Herzegovina, Brazil, Bulgaria, Burkina Faso, Burundi, Cameroon, Chile, China, Colombia, Costa Rica, Cóte d’Ivoire, Croatia, Cyprus, Czech Republic, Denmark, Dominican Republic, Ecuador, Egypt, El Salvador, Estonia, Ethiopia, Finland, France, Georgia, Ger- many, Ghana, Great Britain, Greece, Guatemala, Honduras, Hong Kong, Hun- gary, India, Indonesia, Iran, Ireland, Israel, Italy, Jamaica, Japan, Jordan, Ka- zakhstan, Kenya, Kosovo, Kuwait, Kyrgyz Republic, Latvia, Lesotho, Lithuania, Luxembourg, Macedonia, Malaysia, Mali, Malta, Mauritius, Mexico, Moldova, Mongolia, Montenegro, Morocco, Mozambique, Myanmar, Namibia, Neth- erlands, New Zealand, Nicaragua, Niger, Nigeria, Northern Ireland, Norway, Oman, Pakistan, Panama, Paraguay, Peru, Philippines, Poland, Portugal, Puerto Rico, Romania, Russia, Rwanda, Saudi Arabia, Scotland, Senegal, Serbia, Sierra Leone, Singapore, Slovak Republic, Slovenia, South Africa, South Korea, Spain, Sri Lanka, Sweden, Switzerland, Syria, Taiwan, Tanzania, Thailand, Trinidad and Tobago, Tunisia, Turkey, Uganda, Ukraine, United Arab Emirates, United States, Uruguay, Venezuela, Vietnam, Zaire, Zambia, and Zimbabwe.
Appendix B
Proofs
B1. Considering a Generic Probability Density Function of the Potential Buyer’s Valuation
Assume that Δ/2 potential buyers value x at λ, Δ/2 value it at , and 1 - Δ have a valuation [ , ] distributed according to the probability density function f with cumulative distribution function F. For a rise in Δ to be a mean- preserving spread, E(λ) = λm. Further, γ* maximizes
ˆ
ˆ
( ) (1 ) ( ) ( ) 2
(1 ) (1 ) [(1 ) ] ( )
2
dF
v v dF
Endogenous Property Rights 349
for ˆ and W FE otherwise. Therefore, society’s objective function is strictly concave if
1 * ( ) ( ) 0,
* * *
ˆ ˆ ˆ ˆd vv f f
d
which I assume. For ˆ , a positive γ* is implicitly defined by
ˆ * ˆ ˆ ˆ ˆ1 ( )( ) ( ) ( ) ( ) 0,
* 2(1 )
v v f v F F d
and it rises with Δ since this parameter continues to enter linearly and positively in the first-order condition of society’s problem. For ˆ , the inframarginal effect of a rise in γ is
( )
(1 )[ ( | )] ( ) 2 1
ˆ ˆ ( )
v v E F
and thus for m
ˆ( ) ( | ),v E E its first term is positive and γ* = 1. For v < λm instead, γ* can jump from 0 to the positive solution for Δ sufficiently large if
1
( ) * [ (ˆ ˆ| )] ( ) 0. 2
ˆ 2 2
v F v E F
As in the basic setup, this is the case for α and v small and thus for γ* → 1, ˆ .v Under these restrictions, the left-hand side of the inequality is larger than
( | ) 0. 2 2
v E vv
Q.E.D.
B2. The Political Economy of Property Rights Protection
The first-order condition of society’s problem is
2 2
2
2
( 2 ) ( ) 2 ( ) 0
1( *)
v v l v
for ˆ and
2
m
1 2 (1 )
2 v
l
otherwise, and the inframarginal effect of a rise in γ is
1
( )( 2 ) . 2 2
ˆ ˆ v v l l
350 The Journal of LAW & ECONOMICS
The second-order condition is as in the basic setup. Then γ* is set inefficiently high and may jump from 0 to the positive solution for Δ sufficiently large when
2 2
2 m2 * (1 *)( ) 2 ( ),
* ( )
v v v v
which is easier to satisfy the smaller γ* is for ε not too small and holds for γ* = 1. When high-valuation potential buyers are excluded, society maximizes the function
ˆ
ˆ
(1 ) (1 ) (1 ) (1 )
2
v v d d
l l
for ˆ and
2 FE 1 2
(1 ) 2
W l
otherwise. Hence, the inframarginal effect of a rise in γ and the first- and second- order conditions for a positive solution are as in Section 3.2, whereas for ˆ the first-order condition is
2
m
1 2 (1 )
2 v
l
and thus leads again to an inefficiently large γ*. The condition such that γ* may jump from 0 to the positive solution is easier to satisfy the higher Δ is for
2 2
2
[ * (1 *) 2 2 ] * *
2 ( ) 2
v v v
v v
and thus for ε not too small and sufficiently large. Q.E.D.
B3. More General Market Structures
For ˆ , the first-order condition of society’s problem
2
1 ( )( 2 ) ( )
2 2
1 ( 5 4 ) ( 2 )( 3 2 )
32(1 ) 8
ˆ ˆ
8
v v l
l v v v
l
equals the inframarginal effect of a rise in γ, can be negative only for v < λm , and is concave in Δ. Since for v < λm the left-hand side of the equality is positive for Δ → 1 and can be negative for Δ → 0, it can also change sign
Endogenous Property Rights 351
for Δ large enough. For even larger values of Δ, γ* can jump from 0 to 1 if 2 2 2
m( ) (1 )( 2 ) 2 ,v l v v l which is true (can fail) for Δ → 1 (0), and it is easier to satisfy the larger Δ is since
2 ( )( ) 0.v v
Q.E.D.
B4. The Market for Lemons
A high-valuation potential buyer gains from buying an expected utility of
θ(pL/2) - pL , which is weakly greater than her expected payoff from expro-
priation (1 - γ)θ(pL/2) when γ* is higher than 2/θ < 1 with θ > 2. This last
condition also assures that her expected payoff from buying is weakly positive.
The middle valuation is lower than the high valuation for μ < (θ - 1)v, and
middle- valuation potential buyers do not buy because pL > 2μ entails that pL/2
+ μ - pL ≤ 0. For γ* ≤ 2/θ, the derivative of society’s objective function with
respect to γ is 2
(1 ) ( /4)[( 1) / ]( ), which is always nega-
tive. For γ* > 2/θ, it equals (1 ) ( /4)[( 1)/ ]( ), which is in-
creasing in Δ, negative for Δ → 0, and positive for Δ → 1. As Δ grows bigger
then, γ* can jump from 0 to 1. Social welfare is greater at 1 than it is at 0 for
(1 ) ( /4)[( 1)/ ]( ) 0 and therefore for Δ sufficiently large.
Q.E.D.
B5. Property Rules versus Liability Rules
There are two key observations. First, notice that R
ˆ v since (1 )( ) (1 ) 0.v Second, for R
ˆ R( ˆ ), the derivative of
the social welfare function with respect to γ is
2 2 2
2 2
2
( ) ( )( )
(1 ))
( )
( *
v v l v
2 2
m( ) ( ) ( ) , (2 1 )2
v v v l
which strictly rises with ϕ for every v (v ≥ λm). For R ˆ thus, society’s prob-
lem is convex in γ for 1
1 ( )v and both concave and increasing in
γ otherwise. Since Ψ < 0 for ϕ = 0, there is a 0 < ϕ2 < ϕ1 such that Ψ becomes
positive for ϕ < ϕ2. For R ˆ , Ω < 0 for v < λm since the sum of its first two
terms is at most l(v - λm) < 0 and its third term is negative, and Ω can be posi-
tive for ϕ = 1 and v ≥ λm. For v ≥ λm, since Ω is negative when evaluated at ϕ
= ϕ1 and α not too large, there is possibly a ϕ3 > ϕ1 such that Ω becomes positive
for all ϕ ≥ ϕ3. All in all, γ* is 0 for ϕ < ϕ2 and 1 for ϕ ≥ ϕ3 and v ≥ λm, and it
can jump from 0 to 1 as Δ becomes sufficiently large in the remaining cases. Such
a discrete shift is realized when social welfare is greater at 1 than it is at 0 or
352 The Journal of LAW & ECONOMICS
2 2
2 2 2
m
( ) ( ) (1 ) (1 )
2 2
(1 )( ) (1 ) [ (1 ) ] ,
2
v v v v
l l
v v
l
which is (not) the case for Δ = 1 (0), and it is easier to satisfy the larger Δ is since
2 ( )( ) (1 )( ) 0v v v under my assumptions on α.
Q.E.D.
B6. Heterogeneous (Homogeneous) Original Owners (Potential Buyers)
The first-order condition of society’s problem
2 ( )
(4 3 )ˆ * 2 0 1
v l v v v
is necessary and sufficient, and a positive γ* weakly rises with Δ if the second- order condition 2λ(2λ - 3γ*λ + 2α) is negative. This inequality always holds if it does at γ* = 1—that is, if 2α < λ, which covers here the same role as the re- striction α < min{v, l - v} in the basic setup. Since again the inframarginal effect of a rise in γ—that is,
ˆ ˆ 1
( )( 2 ) ( ) 2 2
v v v v v l
—can be negative only for vm < λ, γ* = 1 if vm ≥ λ. For vm < λ instead, γ* can jump from 0 to the positive solution for Δ large if ( )( * )ˆv̂ v v v
(1 *) ( ) 0ˆ ˆ ,v v v which is true for the lowest interior γ*—that is, 1
( )v —and a fortiori otherwise. Q.E.D.
B7. Private Sales
Potential buyers always prefer the take-or-leave offer to the market price v + α and favor the private sale over expropriation if λ - γv - α ≥ (1 - γ) λ or
P ˆ . For P
ˆ , the first-order condition of society’s problem is
2 2 2 2 2 P P 2
( ) ( ) 2 ( ) 2 0,
1 1 ˆ ˆ
( )*
v l v l v v v
whereas the second-order condition equals 2α2/γ3 > 0 Hence, society’s prob- lem is strictly convex. Thus, γ* is either 1 or the corner solution 0 ( /( )v ) for v < λm ( v ≥ λm), and γ* = 1 is more likely the larger Δ is if ( )( )v v or v since the social welfare evaluated at 0 ( /( )v ) equals λm ( m m[( )/( )]v v ). Q.E.D.
Endogenous Property Rights 353
B8. Costly Expropriation
For C ˆ , the derivative of the social welfare function with respect to γ is
2 2 2
2 2
( ) ,
( *) (1 *)
v c c
whose second term is positive and whose first term decreases with c and is 0 when c is replaced by α < c. For
C ˆ instead, the derivative of society’s objective
function with respect to γ is
2 2
2
1 2 ( ) ,
2 2(1 *)
c v v
l
which rises with γ*. Hence, γ* is either 0 or 1, with the latter prevailing if
2 1
( ) [ ( )( 2 2 )] 0, 2 2
c v c c v l
which is (not) the case for Δ = 1 (Δ = 0 and c not too small) and is eas- ier to satisfy the larger preference diversity is for c not too small or
2 m( )( 2 2 )c c v v . Q.E.D.
B9. Production
The original owner’s expected utility increases with γ* since
2
( )(1 *) ( ) (1 ) (1 ) 0.
2 (
ˆ
*)
v v v v
ll
Thus, rises with Δ if this utility decreases with Δ or m ( )2 1 *v ˆ2 (1 *),v which is true if m
ˆ and so, with * 0,/d d if is suf- ficiently large. Q.E.D.
B10. Investment
When investment enlarges v, the condition ( )/v v implies that (1 ) ,v and the derivative of the original owner’s expected utility
function with respect to I* is
II I
I
*(1 )(1 ) (1 )( ) (1 ) (1 ) (1 ) 0.
*2
ˆ ˆvd v v
d l l
Hence, O rises with Δ if the original owner’s expected utility decreases with it or m I I I*
ˆ( ) *2 1 2 0) ,(1v v which is true if m I ˆ
and so if is sufficiently large. When
OI * , society picks I* instead of γ*—with I* *
and I ˆ ˆ since optimal property rights rise with v—if social welfare is greater
at I* with investment than it is at γ* without investment or
354 The Journal of LAW & ECONOMICS
I I
I I
I
I
ˆˆ
ˆ
ˆ
* *(1 ) (1 ) (1 ) (1 )
(1 *) * * *( *)( ) (1 ) 0.
2
v d d
l l
v v v d
l
When OI* , society chooses O and not γ* if social welfare is greater at O with investment than it is at γ* without investment or
O O
O O
I O
I O
ˆˆ (
ˆ ( )
ˆ
) (1 ) (1 )
(1 ) (1 )
(1 *) * * ( ) (1
( ) ) 0,
2
v d d
l l
v v v d
l
where I ˆ ( O ) is evaluated at O . Society opts for either I* or O , and thus
investment is made if is large enough. Since the inframarginal effect of a rise in γ is as in the basic setup with v(1 + ρ) ( I
ˆ ) in place of v ( ˆ), if v ≥ λm, then γ* = 1 and investment is certain.
When investment shifts the probability mass from the lowest to the highest possible λ, the derivative of the original owner’s expected utility function with respect to γ is
( ) (1 )
(1 ) (1 ) ( ˆ ˆ *
1 ) 0. 2 *
v d v v
l d l
Hence, B rises with Δ if the original owner’s expected utility decreases with it or m I2 (1 *) (1 *) 2 (1 *)
ˆ 0,v v l v which is true if (1 ) (1 ˆ) 2 or for sufficiently large. For B* , society selects
B and not γ* when social welfare is greater at B with investment than it is at γ* without investment or for sufficiently large; that is, when
B B
B m
B
B
ˆ
( )
ˆ
ˆ
( )ˆ
(1 ) (1 ) (1 )
(1
* *
) (1 )
[ (1 ) ]( ) (2* ) 0,
2
v d d
l l
v d
l
v
where ˆ ( B ) is evaluated at B . If B* , society opts for γ* and investment is made. Since the inframarginal effect of a rise in γ is as in the basic setup, if v ≥ λm, then γ* = 1 and investment is certain.
Endogenous Property Rights 355
When investment activities decrease socially wasteful transaction expenses, γ*
maximizes
2
ˆ
ˆ
(1 *) (1 *) (1 )
2
(1 ) (1 ) ( *) (1 )
2 2
d l
v v d
l
for ˆ and W FE otherwise. If ˆ , society’s problem is concave when α(1
- 2ρ*) < v, and its first-order condition equals that in Section 3.2 except for the
investment-inducement term
* 1 1
(1 ) * * * , 2
ˆ ˆ ˆd d d
d l l d l d
which is multiplied by 2l(1 - Δ)–1 when the derivative of society’s objective
function is written as in equation (2). For v < λm , γ* can jump from 0 to
( )( 2 * )
, ( )[ /(1 )]
v v
v v l
which is higher than the property rights level prevailing without investment and
rises with Δ, for Δ large enough. This is the case if
2 2
m
m
2 [ * (1 *)] 2 * *
* 2 ( ) *
ˆ
2 ( ,ˆ )
v v
d v v
d
which for m ˆ is easier to satisfy the smaller ρ* is and for ρ* = 1 is always
true. Since the inframarginal effect of a rise in γ is the same as in the basic setup,
γ* = 1 for v ≥ λm. Q.E.D.
References
Acemoglu, Daron, and Simon Johnson. 2005. Unbundling Institutions. Journal of Political
Economy 113:949–95.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. The Colonial Origins
of Comparative Development: An Empirical Investigation. American Economic Review
91:1369–1401.
———. 2002. Reversal of Fortune: Geography and Institutions in the Making of the Mod- ern World Income Distribution. Quarterly Journal of Economics 117:1231–94.
Aghion, Philippe, Alberto Alesina, and Francesco Trebbi. 2004. Endogenous Political In-
356 The Journal of LAW & ECONOMICS
stitutions. Quarterly Journal of Economics 119:565–611.
Aghion, Philippe, Yann Algan, Pierre Cahuc, and Andrei Shleifer. 2010. Regulation and
Distrust. Quarterly Journal of Economics 125:1015–49.
Ahlerup, Pelle, and Ola Olsson. 2012. The Roots of Ethnic Diversity. Journal of Economic
Growth 17:71–102.
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain
Wacziarg. 2003. Fractionalization. Journal of Economic Growth 8:155–94.
Altonji, Joseph G., Todd E. Elder, and Christopher R. Taber. 2005. Selection on Observed
and Unobserved Variables: Assessing the Effectiveness of Catholic Schools. Journal of
Political Economy 113:151–84.
Ashraf, Quamrul, and Oded Galor. 2013. Genetic Diversity and the Origins of Cultural
Fragmentation. American Economic Review 103:528–33.
Bar-Gill, Oren, and Nicola Persico. Forthcoming. Exchange Efficiency with Weak Prop-
erty Rights. American Economic Journal: Microeconomics.
Barry, Jordan M., John William Hatfield, and Scott Duke Kominers. 2014. Coasean Keep-
Away: Voluntary Transaction Costs. Unpublished manuscript. Harvard Business
School, Boston.
Bejtja, Saida, and Dritan Bejtja. 2014. Private Property Issues on Eastern Europe in Res-
titution and Compensation Problems. Academic Journal of Interdisciplinary Studies
3:271–78.
Bell, Abraham. 2009. Private Takings. University of Chicago Law Review 76:517–86.
Bellows, John, and Edward Miguel. 2009. War and Local Collective Action in Sierra Le-
one. Journal of Public Economics 93:1144–57.
Besley, Timothy, and Maitreesh Ghatak. 2010. Property Rights and Economic Develop-
ment. Pp. 4525–95 in vol. 5 of Handbook of Development Economics, edited by Dani
Rodrik and Mark Rosenzweig. Amsterdam: Elsevier.
Besley, Timothy, and Torsten Persson. 2009. The Origins of State Capacity: Property
Rights, Taxation and Politics. American Economic Review 99:1218–44.
Boranbay, Serra, and Carmine Guerriero. 2015. Endogenous (In)Formal Institutions. Un-
published manuscript. University of Amsterdam, Amsterdam Center for Law and Eco-
nomics, Amsterdam.
Bouckaert, Boudewijn, and Gerrit De Geest. 1995. Private Takings, Private Taxes and
Private Compulsory Services: The Economic Doctrine of Quasi Contracts. Interna-
tional Review of Law and Economics 15:463–87.
Calabresi, Guido, and A. Douglas Melamed. 1972. Property Rules, Liability Rules and In-
alienability: One View of the Cathedral. Harvard Law Review 85:1089–1128.
Cavalli-Sforza, Luca L., Paolo Menozzi, and Alberto Piazza. 1994. The History and Geogra-
phy of Human Genes. Princeton, NJ: Princeton University Press.
Cervellati, Matteo, Piergiuseppe Fortunato, and Uwe Sunde. 2008. Hobbes to Rousseau:
Inequality, Institutions, and Development. Economic Journal 118:1354–84.
Charnoz, Olivier, and Ashwini Swain. 2012. In Pursuit of Energy Efficiency in India’s Ag-
riculture: Fighting “Free Power” or Working with It? Working Paper No. 126. Agence
Française de Développement, Paris
Cook, R. Dennis. 1977. Detection of Influential Observations in Linear Regression. Tech-
nometrics 19:15–18.
Dari-Mattiacci, Giuseppe, and Carmine Guerriero. 2015. Law and Culture: A Theory of
Comparative Variation in Bona Fide Purchase Rules. Oxford Journal of Legal Studies
35:543–74.
Endogenous Property Rights 357
———. 2016. A Novel Dataset on Horizontal Property Rights in 126 Jurisdictions. Un- published manuscript. Amsterdam Center for Law and Economics, Amsterdam.
Dari-Mattiacci, Giuseppe, Carmine Guerriero, and Zhenxing Huang. 2016. The Property-
Contract Balance. Journal of Institutional and Theoretical Economics 172:40–64.
Desmet, Klaus, Michel Le Breton, Ignacio Ortuño-Ortín, and Shlomo Weber. 2011. The
Stability and Breakup of Nations: A Quantitative Analysis. Journal of Economic Growth
16:183–213.
Diamond, Peter, and Johannes Spinnewijn. 2012. Capital Income Taxes with Heteroge-
neous Discount Rates. American Economic Journal: Economic Policy 3:52–76.
Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2003.
Courts. Quarterly Journal of Economics 118:453–517.
Fischel, William A. 1995. Regulatory Takings: Law, Economics, and Politics. Cambridge,
MA: Harvard University Press.
Glaeser, Edward L., and Andrei Shleifer. 2003. The Rise of the Regulatory State. Journal of
Economic Literature 41:401–25.
Guerriero, Carmine. 2016. Endogenous Legal Traditions and Economic Outcomes. Jour-
nal of Comparative Economics 44:416–33.
Gwartney, James D., and Robert Lawson. 1997. Economic Freedom of the World: 1997 An-
nual Report. Washington, DC: Heritage Foundation.
Hall, Bronwyn H., and Dietmar Harhoff. 2012. Recent Research on the Economics of Pat-
ents. Annual Review of Economics 4:541–65.
Hart, Oliver, and John Moore. 1990. Property Rights and the Nature of the Firm. Journal
of Political Economy 98:1119–58.
Hasen, Richard L., and Richard H. McAdams. 1997. The Surprisingly Complex Case
against Theft. International Review of Law and Economics 17:367–78.
Johnson, Simon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2000.
Tunneling. American Economic Review 90:22–27.
Jordan, Jim. 2006. Pillage and Property. Journal of Economic Theory 131:26–44.
Kaplow, Louis, and Steven Shavell. 1996. Property Rules versus Liability Rules: An Eco-
nomic Analysis. Harvard Law Review 109:713–90.
La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny. 1999.
The Quality of Government. Journal of Law, Economics, and Organization 15:222–79.
Lieberman, Evan S., and Gwyneth McClendon. 2013. The Ethnicity-Policy Preference
Link in Sub-Saharan Africa. Comparative Political Studies 45:574–602.
McEvedy, Colin, and Richard Jones. 1978. Atlas of World Population History. New York:
Facts on File.
Michalopoulos, Stelios. 2012. The Origins of Ethnolinguistic Diversity: Theory and Evi-
dence. American Economic Review 102:1508–39.
Murray, Damian R., and Mark Schaller. 2010. Historical Prevalence of Infectious Diseases
within 230 Geopolitical Regions: A Tool for Investigating Origins of Culture. Journal of
Cross-Cultural Psychology 41:99–108.
Nicita, Antonio, and Matteo Rizzolli. 2014. In Dubio Pro Reo. Behavioral Explanations of
Pro-defendant Bias in Procedures. CESifo Economic Studies 60:554–80.
Olsson, Ola, and Christopher Paik. 2016. Long-Run Cultural Divergence: Evidence from
the Neolithic Revolution. Journal of Development Economics 122:197–213.
Piccione, Michele, and Ariel Rubinstein. 2007. Equilibrium in the Jungle. Economic Jour-
nal 117:883–96.
Pinhasi, Ron, Joaquim Fort, and Albert J. Ammerman. 2005. Tracing the Origins and
358 The Journal of LAW & ECONOMICS
Spread of Agriculture in Europe. PLoS Biology 3:2220–28.
Putterman, Louis, and David N. Weil. 2010. Post-1500 Population Flows and the Long-
Run Determinants of Economic Growth and Inequality. Quarterly Journal of Economics
125:1627–82.
Sackman, Julius, Russell van Brunt, Patrick J. Rohan, Melvin Reskin, Tony Prince Brigham,
and Gideon Kanner. 2013. Nichols on Eminent Domain. 3rd ed. Albany, NY: Matthew
Bender.
Saez, Emmanuel. 2002. The Desirability of Commodity Taxation under Non-linear In-
come Taxation and Heterogeneous Tastes. Journal of Public Economics 83:217–30.
Segal, Ilya, and Michael Whinston. Forthcoming. Property Rights and the Efficiency of
Bargaining. Journal of the European Economic Association.
Williamson, Oliver E. 2010. Transaction Cost Economics: The Natural Progression. Amer-
ican Economic Review 100:673–90.