Voodoo and fishing in a changing world: The role of traditional religion and modern

advertisementhumphΔιαχείριση

9 Νοε 2013 (πριν από 4 χρόνια και 4 μέρες)

247 εμφανίσεις

Voodoo and fishing in
a changing world
:
The role of traditional religion and modern
institutions in managing the

inland
fisheries

of Benin
1

Elena Briones Alonso
2

Marijke Verpoorten
3


Romain Houssa
4


This draft:
June 2013

Abstract

This paper investigates the
role of traditional and modern community
-
level institutions
in
managing

the inland fishery stock of southern Benin.
More specifically, w
e
investigate

the
functioning of two rules

de
signed to regulate the use of fine mesh

fishing
nets.

The rules stem
from two distinct institutions
:

(i)
a

modern
institution embodied in fishing committees
,

and
(
ii)
a

traditional institution

embedded in
Voodoo
,
an
animist
ic

religion that originated in
Benin
.

We examine
whether these rules are respected,

relying on individual
-
level data on the
use of fishing
gear
,

fishing revenue

and religious affiliation.

We

find that
the traditional and
the modern rule co
-
exist: both rules
have a statistically significant effect on fishing
activities
.

However,
the
ir

quantitative effect is small

and, unless reinforced,

the existing rules

may not
be sufficient to
prevent
a

collapse of the ecosystem.


Key words:
common
-
pool
resource
management;

fisheries;

governance
institutions
;

social
capital;
A
nimism
;

Voodoo;
Benin

(
West
-
Africa
)




1

This research was funded by the Centre for Institutions and Economic Performance (LICOS) and the Fund for
Scientific Research


Flanders (FWO
-
Vlaanderen). We would like to thank
the seminar
participants of the
May
2012
CRED
-
LICOS Workshop
, the 2012 Confer
ence ‘Design and Dynamics of Institutions for Collective
action’ at Utrecht University and the February 2013 Economics Seminar at Wageningen University
. The usual
disclaimer applies.

2

University of Leuven (KU

Leuven)



Centre for Institutions and Economic Performance (LICOS)

3

University of Antwerp (U
A)
-

Institute for Development P
olicy and Management (IOB) and University of
Leuven (KU

Leuven)


Centre for Institutions and Economic Performance (LICOS)

4

CRED & CeReFiM,

University of Namur.
Tel: +32(0)81 724947. Email: romain.roussa@fundp.ac.be



1.

INTRODUCTION

Around

the globe, marine and inland fisher
y stocks are being overexploited

(
Allan et al.,
2005;
Food and Agriculture Or
ganization
,

2012)
.
According to Hardin (1986),
overexploitation of
common pool r
esources such as fisheries
is inevitable unless the resource

is

converted to state or private property
.

As Hardin theorizes
, this inexorable ‘tragedy of the
commons’ rises from the fact that a
ll
common pool

resources are characterized by
non
-
trivial
exclusion and
rival
ry

in consumption. These properties create

short term

individual

incentive
s

for

unlimited

exploitation

that conflict

with the long
-
run interest of the group to restrain
exploitation
and preserve the resource
. The argument that this
social dilemma

will inevitably
lead to overexploitation of the resource hinges on the assumption

that resource users are
unable to collect
ively restrain

exploitation

without the intervention of an external
force.


An extensive body of theoretical and case study literature has shown that this
assumption
does not always hold
:

c
ommunities

of resource users
often succeed in

collectively regulating
exploitation and sustainably managing common pool resources

(Baland & Platteau, 1996;
Berkes, 1989; Berkes et al, 1989; Bromley, 1992; Feeny et al., 1990; Feeny, Hannah &
McEvoy, 1996; Gibson et al., 2000; Pinkerton & Weinstein, 1995; McCay & Acheson, 1987;
National Research Council [NRC], 2002; Ostrom, 1990; Wade
, 1994).

The

success of
community
-
level governance
is shown to crucially
depend
on the existence of effective
governance institutions
, which has led to the formulation of

design principles for
effective
institutions (Agrawal, 2001; Cox, Arnold & Villamayor

Tomás, 2010;
;
NRC, 2002; Ostrom,
1990
; 1999
; Wade, 1994)
.


Several design principles tend to be met when
resource governance institutions are embedded
in
traditional
culture and
religion
,

and the historical success of this type of institutions
support
s

the notion that
traditional
religion
can serve as an
adequate

framework for
governance institutions

(
Barua, 2009;
Berkes, Colding & Folk, 2000;
Dorm
-
Adzobu,
Ampadu
-
Agyei
&

Veil
, 1991
;
Kajembe, Luoga, Kijazi & Mwaipopo, 2003;
Nyamweru &
Sheridan, 2008;
Rusinga & Maposa, 2010; Sharma, Rikhari & Palni, 2010).
However,
it was
also shown that traditional

institutions

are often vulnerable to changes in the environment
, in
particular when these changes take the form of the arrival of new, competing governance
institutions
(
Baland & Platteau, 1996;
Dorm
-
Adzobu, Ampadu
-
Agyei & Veil, 1991;
Eneji et
al., 2012)
. Adaptability to a changing environment thus emerged as a crucial principle in the
design of
effective

governance institutions (
Dietz, Ostrom & Stern

2003; N
RC
, 2002).

In this
context an important question ari
se
s
:
to what extent
can traditional

religion
continue to serve

as a framework for
effective

governance institutions

in an increasingly globalized and
modernized world
?


We examine this issue in the context of the coastal fisheries of
s
outhern Benin, which provide
a livelihood
to

several communities of artisanal fishers
, yet are severely affected by natural
resource degradation
. More specifically, we
study

the effectivenes
s of
two institutions
designed to manage the fishery stock. One is a traditional pre
-
colonial institution integrated in
Voodoo, an

a
nimist
ic

religion
that
originated

in Benin. The second is a post
-
colonial
institution in the form of rules issued by fishing committees (
comités de pêche
), a relatively
new structure.
Both institutions formulate rules about the use of
fine

mesh
fishing
nets
(

konou

),
but the rules differ
.
Whereas Voodoo prohibits the

use of fine

mesh nets
at all
times
, the fishing committees prohibit the
use of the
konou

only
periodically.
We are thus
presented with a

case in which two
competing

i
nstitutions exist side by side
:

one originating
from
traditional religion in
pre
-
colonial times and one
secular institution
created in recent
years.



In

contrast to

many other
cases of competing institutions,
the modern institution was not
created and imposed by an external force, such as the central govern
ment, but created by the
fishing

communities themselves.
The
local fishing committees were
formed
in response to
problems of overfishing and resource degradation
,

with the task of regulating the use of
harmful fishing gear and in particular the
use of the
konou
.

The committees decided upon a
rule that applies throughout the fishing season and allows the use of the
konou

for four
consecutive weeks

(‘open’ weeks)
, then bans it for the following two weeks

(‘closed’ weeks)
.


The

traditional Voodoo
institution

does not stipulate any rules about
the
use

of the
konou

specifically, as it is a relatively new fishing instrument (initiated in 1986)
. There are,
however, traditional rules about the use of fishing gear

with fine mesh nets
, and these differ
from the mode
rn rule. Instead of allowing the use of fine mesh
fishing
nets in some periods
and abolishing it in others, the Voodoo
rule

bans the use of fine mesh
nets

at all times
(Dangbégnon, 2000; Clédjo, 2006; Pliya, 1980).


In this
study,

we seek to address two s
pecific questions. First, does Voodoo


which in its
heydays successfully banned the use of fine mesh nets


still induce fishermen to refrain from
using fine mesh fishing gear?

And second,
is the modern rule effective in regulating the use
of the
konou
?
A
nswering

these questions allows us to assess whether traditional religion
-
based
institutions

can
continue to influence
behaviour

in
a

changing
environment
,
and
whether new locally devised institutions can be effective in this setting.

Benin provides an ideal testing ground for the first question, because of its unique setting of
religious pluralism, with considerable variation in reported religion between and within
villages and even within the same household (Barbier & Dorier
-
Apprill,
2002). This
religious
heterogeneity
not only increases the probability that reported religion in Benin reflects
individual beliefs, but also allows us to measure the impact of religion on fishing
behaviour
after controlling for other relevant individual an
d community characteristics. In contrast,
most studies so far have studied homogeneous religious groups and
have
therefore faced the
difficulty of perfect correlation in these variables.



We explore the correlation between Voodoo and the use of the
konou

in two different
samples.
Both samples contain data on fishers living at Lake Nokoué and Lake Porto Novo,
two of the main coastal lakes in Benin.
The first is a sample of 392 fishermen
included
in

our
household
survey for which we have observations across

several weeks in 2009. The second
is the 2006 fishery census, which contains one single observation for

18,683 fishermen from
all 9 communes surrounding Lake Nokoué and Lake Porto Novo. Using the large census
dataset allows us to better control for enviro
nmental factors


by means of village fixed
effects


that may confound the relationship between Voodoo and the choice of fishing gear.

If
the
Voodoo
rules are
still
respected
, we should find that Voodoo followers use
fishing gear
with fine mesh nets signi
ficantly less, i.e. Voodoo followers should
have a lower probability
of using the
konou
.


To investigate
the effectiveness of

the modern rule
, we rely on a detailed dataset of the self
-
reported use of fishing gear and fishing revenue over time, for a perio
d of 14 weeks. If fishers
respect the modern rule, we should find a corresponding periodical variation in the use of the
konou

across weeks. Fishermen may, however, lie about their use of the
konou
. If they do, the
variation we pick up could merely reflect clever but untruthful answering by the respondents.
To
verify

whether

our results are driven by untruthful answering, we examine the fluctuations
in fishing revenue reported by fishers. If fishers

respect the modern rule, we should find
fishing revenue to be lower in closed weeks. Moreover, if the use of the
konou

is abandoned
during the closed weeks, the resulting fall in fishing intensity should allow the fishery stock
to increase. Fishing revenu
e should then be highest in the first week of re
-
opening, and
gradually decline in the following open weeks as intensive fishing with the
konou

again
reduces the fishery stock. Finding such a trend across open weeks would provide additional
evidence that t
he modern rule is respected.


Relying on our

quantitative analysis, underpinned with insights from additional qualitative
interviews
, we provide evidence that both the traditional and the modern
rule

affect
the use of
the
konou
, but that both are malfuncti
oning as their impact is limited.
First, for
both our
sample of fishermen and the fishery census of 2006 we find a negative
relation

between
adherence to Voodoo and the use of the
konou
, suggesting that the traditional Voodoo rules
are still respected by at least some fishermen.
Second, t
he variation in the use of the
konou

and
in
fishing revenue across weeks
suggest
s

that the
konou

is less used when the lakes are
closed, but that compli
ance to the modern rule is far from perfect.
We explore a number of
competing explanations for these results and find that our qualitative conclusions hold.



The next section

gives an overview of
the coastal fisheries of Benin and the

problem of
overfish
ing, and discusses the evolution of the fishery governance institutions
. In section 3

we set out the empirical framework used to address our research questions.
Section 4
discusses

the data
used in the analysis
and section 5 presents the results.

Section 6

concludes.


2. THE COASTAL FISHERIES OF BENIN


2.1.
Natural resource system

and
degradation

W
e
study fishing communities

located

at
two of the largest lakes in
s
outhern

Benin
:

Lake
Nokoué and

Lake

Porto Novo
. Both lakes are

indicated on the map in Figure
1
.

Lake Nokoué
is by far
the largest water body, with a surface of 150 km
2
, while L
ake Porto Novo is
considerably smaller (35 km
2
) and shallower (Gnohoussou, 2006). The two lakes are part of
the largest and most productive water basin in Benin


responsibl
e for 65 to 70 percent of the
total inland fisheries production


and are connected via the Totché channel (Gnohoussou,
2006;
Clédjo, 2006
). Lake Nokoué is
also

connected to the Atlantic Ocean through the
Cotonou channel
, which

is an important element in t
he

lake

ecosystem: the periodical inflow
of marine water creates seasonal variations in the salinity and temperature of the
lake
s
, which
promotes the diversity, reproduction and growth of aquatic fauna and flora (Amoussou, 2004;
Gnohoussou, 2006
; see
Cummings (1961) and Williams (1958) for the case of southern pink
shrimp).


In the course of history different ethnic groups settled around the lakes and engaged in
fishing activities (Bourgoignie, 1972; Pliya, 1980; 1989). The lakes gradually became
surro
unded by fishing villages, located on the lake shores or on the water in the form of pile
villages (Bourgoignie, 1972; Pliya, 1980; 1989). To this day, only artisanal fishers are active
on the lakes; industrial fishing is absent. The livelihoods of the loc
al households are highly
depe
ndent on the lake fisheries:
there are few alternative economic activities available to
them
, which

is
reflected in a very low degree of

income diversification (Stoop et. al, 2012
).


T
he

coastal
lakes
suffer from

severe environ
mental degradation and overfishing (Amoussou,
2004; FAO, 2008; Gnohoussou, 2006; USAID, 2007). The main causes are strong population
growth
, commercialization
and monetization of the economy,
pollution from industrial and
household waste

and the erosion of

the traditional Voodoo governance system

(Amoussou,
2004; FAO, 2008; Gnohoussou, 2006; USAID, 2007). The increasing value of fish
ery
products

led

to
a rise in

fishing intensity and triggered the search for more powerful fishing
technologies (Dangbégnon, 2
000). The most important innovations in fishing techniques
were the
acadja
, introduced around the 1960s, and the
konou
, introduced around the 1980s
5
?
?


T
he
konou

(also known as

medokpo
konou
)

is a fixed fishing installation consisting of a
central rectangular fine mesh net (20 to 5 mm)
with
pouches at
its

extremities (USAID, 2007;
PNE, 2010).
Animals that hit the central net are guided into the pouches, from which they
cannot escape.
This struc
ture allows

the
konou

to
catch
large
quantities

of fish
and shrimp
,
especially

when placed in narrow channels or strong currents.
Although the
konou

is not
designed exclusively for catching shrimp,
it is one of the main
drivers of fishing revenue
from shri
mp fishing
6
?
?
7KH?
konou

is also rather expensive
compared to other fishing gear used
at these lakes

The
acadja

is a type of tropical brush park fishery that was developed around the end of the
19
th

century and modernized by the government in the 20
th

century. It is constructed by
placing wooden branches in the bottom of the lake and fenci
ng these branches with a fishing

net. The branches, which resemble the natural mangrove habitat of several fish species,
protect spawning areas and
provide fish with
shelter and abundant food supply (de Kimpe,
1968; Lalèyè, 2000; Niyonkuru & Lalèyè, 2010; Gnohoussou, 2006). These
features

attract
fish to the
acadja

and

stimulate fish reproduction

and growth

within the
acadja
,
making this

a
highly productive fishing ins
trument (
de Kimpe, 1968;
Clédjo, 2006; Lalèyè, 2000;
Niyonkuru & Lalèyè, 2010).





5

Other fishing gear used
on the lakes
include
cast nets, trawl nets, pots, long

lines, hooks and lines, landing nets
and fish traps.

6

The other main driver of fishing revenue from shrimp fishing is the shrimp pot (nasse à crevettes).

Over the last decades, t
he proliferation of
acadjas

has been

a major cause of the
disappearance of mangrove vegetation on the lake shores, which has aggravated the problems
of siltation and habitat degradation (Lalèyè, 2000; Gnohossou, 2006; USAID, 2007). Today,
the
net environmental impact of the use of the
acadja

is debated. There is
general
agreement
that
it

increase
s

the amount of organic biomass in the water,
thereby
lowering

oxygen levels
and
leading

to eutrophication (Gnohossou, 2006;
Clédjo, 2006
). Some authors argue that the
presence of
acadjas

itself leads

to siltation, though others claim that this effect is negligible
(see Ammoussou, 2006; Gnohoussou, 2006;
Clédjo, 2006
). Clédjo

(2006) and Niyonkuru and
Lalèyè (2010) argue that the
acadja
, when properly managed, could provide a viable and
sustainable fish
ing practice by promoting fishery reproduction and growth.

The negative impact of the
konou
, in
contrast, is widely agreed upon.

The damaging nature of
the
konou

lies in the use of fine mesh
fishing
nets. These nets also catch non
-
mature animals
and eggs, which reduces the reproduction capacity of the fishery stock and thus endangers the
sustainability of the resource (PNE, 2010; USAID, 2007). In this manner the intensive use of
the
konou

has contribu
ted to the problem of overfishing and resource degradation (PNE,
2010; USAID, 2007).


Aside from environmental concerns, the
acadja

and the
konou

create social tensions as well.
In

the

last

years, the use of these instruments has provoked se
vere conflicts

among fishermen
(Pliya, 1980; Dangbégnon, 2000; PNE, 2010).

The lake and its resources have traditionally
been collective property, open to all fishers. With the introduction

of the
acadja
,

fishers
began to claim exclusive and transferable property rights to parts of the lake surface and
fishing grounds (Pliya, 1980;
PNE, 2010)
.
As
acadja
s

were installed by the first
-
come first
-
serve principle and are transferred from father to son, those fis
hers that do not own an
acadja

are being denied access to an ever
-
growing part of the lake.
The use of the
konou

creates
conflicts because of the negative impact on fishery stocks, and because the
konou
s

are set
in

such ways that they leave few catches to
those fishing in the
neighbo
u
rhood

(PNE, 2010)
.



2.2.
Governance

of fishery resources

The traditional Voodoo institution

In pre
-
colonial times, the fishing activity
at

the
southern
lakes

of Benin was regulated by the
traditional animistic religion of Benin: Voodoo (Bourgoignie, 1972; Pliya, 1980).

In the
Voodoo religion, Voodoo spirits each have their o
wn task in ruling the world and
control
those elements in the physical wo
rld to which they are connected.
For instance, the Voodoo
spirits that govern the lakes are believed to control
the movements of water and fish.
Voodoo
spirits can thus protect and help mankind, but can also do
harm

when angered or offended
.

Phenomena such

as flooding or
death by drowning

are believed to be punishments f
rom an
angered Voodoo spirit
.

However, Voodoo spirits in turn need mankind to nurture and sustain
their power
.
T
he Voodoo

belief system
thus creates

a strong interdependence and integration
of the human and the divine

(Bourgoignie, 1972; Tall, 1995a).

At Benin’s southern lakes
,
the institution governing the lake fisheries was embedded in the
Voodoo religion.

A wide array of concrete
rules

and taboos
integrated in Voodoo beliefs
regulated
fishing

intensity and the use of fishing gear

with the aim of

preventing

overfishing
and
preserving

the collec
tive nature of the resource

(Bourgoignie, 1972; Dangbégnon, 2000;
Clédjo, 2006; Pliya, 1980)
. Fishing was for instance banned during market days a
nd days of
Voodoo worship (in practice two days per week
)
, and

was prohibited in the vicinity of sacred
locations,
which

often

coincided with spawning grounds for fish.
Rules regarding fishing gear

for instance
prohi
bited the use of fine mesh nets and

the use of multiple hooks (Dangbégnon,
2000;
Clédjo, 2006
; Pliya, 1980).

R
ules were organized and enfor
ced by the local spiritual
leaders (Voodoo priests),
who gained legitimacy, authority and trustworthiness from their
politico
-
religious status (Bourgoig
nie, 1972; Dangbégnon, 2000; Clédjo, 2006; Pliya, 1980).

Respected and feared

because of their
connect
ion

to
Voodoo
spirits
and their alleged
witchcraft
powers
, Voodoo priests played a crucial role in the organization, monitoring and
enforcement of the tra
ditional
Voodoo
rules (Dangbégnon, 2000; Pliya, 1980).
Sanctions
were severe, ranging from fines to public flagellation or the confiscation of fishing gear.

T
he

worst offenses
were

sanction
ed by

the death sentence, which was depicted as a punishment
by the

offended Voodoo spirit and executed by

Voodoo p
riests (Pliya, 1980).



According to Pliya (1980), the Voodoo
governance
institution

managed to keep resource
exploitation in check, even in the face of population growth.
The

system started to fail
,
however,

when
the colonization of Benin brought about
some profound socio
-
economic
changes

(Dangbégnon, 2000; Pliya, 1980)
. The

traditional
politico
-
religious
structures

were
undermined

by new colonial and post
-
colonial powers
, and these
new

powers
also
introduced

monotheistic religions that
started to
compete with Voodoo
.
As the power of Voodoo
declined
, t
he

deterring effect of sanctions decreased
. A
t

the same time
,

the benefits of
shirking increased
with the rising

value of fishery products, resulting from the
monetization
and commercialisation
of the economy and access to new markets
(
urban
centres

and

exports
to Europe
)
.
The economic opportunities created by a growing fishing sector, combined with a
booming population, brought about large flows of internal mig
ration to the southern lakes.
Newly settled a
gricultural communities started exploiting the lake resources as well,
engaging
in

a

competition with the communities who had been full
-
time fishers since pre
-
colonial times. These part
-
time fishers showed littl
e respect for the traditional Voodoo
system, fishing whenever and wherever they chose, openly disobeying rules and
undermining

the au
thority of Voodoo priests. The influx of

outsiders thus further
eroded

the power of the
Voodoo system and reduced the incen
tives to obey the rules (Pliya 1980).


The
waning power of the

traditional Voodoo institution led to an institutional vacuum, which

the

Beninese

governme
nt attempted to fill by creating new governance institutions
. Yet, these
government
institutions faile
d to effectively regulate fishing activities (Maarleveld &
Dangbégnon, 1999; Pliya, 1980).
R
ules were left unmonitored, sanctions were
too lenient and
punishments
were rarely and inconsistently
implemented. For instance, civil servants
designated to
inspect the use of fishing gear

explained in interviews that
monitoring and
certainly sanctioning
are halted
when elections
approach
, in an attempt by the incumbent to
win the votes of the fishers

(see also Dangbégnon, 2000)
. Besides failing to effectively

regulate the fishing activity, the governmentally created institutions
further undermined

the
authority of Voodoo priests

and the traditional rules that they represented

(Pliya
,

1980). For
example, by enforcing the individual property claim of
acadja

owne
rs, the government broke
with the long
-
established Voodoo
principle

that
pre
se
rv
ed

the
lake

and its resources
as
common property (Pliya, 1980).


In spite of these undermining forces, the Voodoo governance system did not completely
disappear. Although the religious landscape in post
-
colonial Benin was dominated by the
Catholic Church, Voodoo remained influential and even adapted to the changed socio
-
economic environment through the emergence of new Voodoo cults (Tall, 1995a; 1995b).
The religion

came under serious pressure
under the Marxist
-
Leninist regime
in Benin
(1972
-
1989)
, which

actively
targeted Voodoo with its anti
-
religious campai
gns and laws

against
witchcraft. However, the tide turned with the

end of the communist regime and the
consequent democratic renewal
.

Authorities restored t
he V
oodoo
religion and actively
supported it, promoting Voodoo as a symbol of

national identit
y and African/Beni
nese
culture
(Tall, 1995b).

Important symbols of the
new

attitude towards Voodoo were the
organization of a national Voodoo

conference

and the enlistment of Voodoo in the
constitution
as an official religion
(Tall, 1995b).
As a consequence, the

Vo
odoo
religion
regained vitality and
became more and more

organized

as a traditional religion, with a
national
festival

(10 January) and a national hierarchy

of Voodoo spirits
(Pliya, 1980; Tall,
1995
b
)
.


Today,
Voodoo
continues to
be deeply intertwined with

day
-
to
-
day life

in Benin
,

and the link
between Voodoo and

fishing activities
remains
(Amoussou, 2004; Clédjo, 2006)
.

Voodoo
priests
are still

involved in the fishing activity

and



alt
hough not all tab
oos and sanctions
have survived


concrete rules and sanctions still exist today (Amoussou, 2004;
Clédjo, 2006
;
Dangbégnon, 2000
).
One rule that has survived and seems to be quite powerful today is the
taboo to

fish

in the neighbo
u
rhood of
Voodoo shrines (
fétiches
)
(
Clédjo, 2006
).
This tab
oo

seems to be
well respected by
fishers


even those who follow another religion


and is
strongly associated with Voodoo beliefs
. The reason for respecting this taboo, as declared by
the fishermen, is a fear of punishment by the offended Voodoo spirit.
It is believed that the
Voodoo spirit will take away the fishing gear of the offender, and in case of repeated
violation of the taboo, may even cause the offender to drown. Since the location of the
Voodoo shrines is usually very visible and always known b
y the local fishers, community
monitoring takes place. Whenever someone is seen fishing in the sacred area, fishers will
warn him of the taboo and the dangers of disobeying.

Another example of
a traditional
Voodoo rule that is still known today is the tabo
o of women going on the lake when
menstruating. Fishermen state that this taboo ex
ists because it is considered dis
respectfu
l

to
bring Voodoo spirits in contact with menstruating women, and this act will undermine their
power to pr
ovide large quantities of

fish. This reasoning again indicates that the fishing
activity continues to be associated with Voodoo spirits and beliefs today, and at least some
traces of the traditional Voodoo governance system remain.

The institution of the fishing committees

As

was

highlighted in the introduction, fishers recently created a new rule to regulate the use
of the
konou
, which is unrelated to the

traditional Voodoo institution
.
This modern
rule is
implemented
a
cross the whole of lake Nokoué and
throughout

the fishing season,
which
coincides with high spawning activities of targeted species (Pliya, 1980). In this period, the
size

of
the
sexually mature
stock has an important impact on

the reproduction rate

and, in
addition, larvae and young specimen have hi
gh growth rates (see Cummings, 1961; Williams,
1958).
By
periodically closing the lakes to

the use of the
konou
, the rule intends to promote
fishery reproduction

and growth

and reduce the damaging impact of the
konou
.

The
committees impose a number of sanc
tions when the rule is violated, such as the confiscation
of fishing gear

or catches when these include young fish.

However,
it appears that
the
effectiveness of the sanctioning mechanism
is undermined by corruption
(Dangbégnon,
2000
).

Monitoring is perfor
med by the members of the fishing committees, but
community
monitoring
seems to take place to some extent as well. In general it appears that
this rule
regarding the use of the
konou

is well known by
fishermen

active at lake Nokoué
.


3.

EMPIRICAL
SETUP

We investigate the functioning of
two institutions regulating the fishing activity. The first is
the
modern rule, which
periodically
prohibits

the

use of the
konou

across weeks
.

T
he
second
is the
traditional
Voodoo
rule, which prohibits the use of
fine
mesh

nets

at all times
. To
identify
compliance to
the modern rule, we exploit variation in the use of
the
konou

across
weeks.
To

analy
s
e

the functioning of
the traditional rule
we exploit
variation in
Voodoo

adherence across and within villages. Our first
equation

to be estimated
is the following
:





































(1)




is an indicator variable taking value 1 if individual
i

reports to use the
konou

for
fishing in week

t
, and 0 otherwise;




is an indicator variable that equals 1 if the lakes
are

closed in week
t
;




is an indicator variable
that equals

1 if individual
i

reports his
religion

to be

Voodoo
;




is a time trend;





is a vector of the following
control
variables for
individual
i

in week
t
:

the logarithm of age, an indicator variable for literacy
, an interaction
effect between

literacy and



and the logarithm of the total value of assets;



indicates
fixed effects at the
arrondissement

level
7
;



denotes the usual error term.

The time trend is included to capture temporal variation in the use of the
konou

that is
generated by
unobserved
environmental
variables
, most importantly the
growth cycle

for
shrimp in the survey period March
-
June
. We control for age
because

the use of the
konou

requires considerable physical strength. Literacy and the total value of assets are i
ncluded a
s
indicators of
(permanent)
income
,

which may play a role as
it

is rather expensive
to
acquire

the
konou
.
The

ex
clusion
of these variables leaves our results basically unaffected,
attenuating possible endogeneity concerns.
The interaction term between lite
racy and the
closed variable is included to capture differences in compliance between fishermen of high
and low education levels.


If




is estimated significantly

negative
,
it

suggests that the use of the
konou

is on average
lower for weeks in which
the modern rule prohibits its use
, i.e. for weeks in which the lakes
are

closed

. If



is estimated significantly negative, it suggests that V
oodoo

fishermen

respect the
traditional
rule of not using
fine mesh
ed nets.
However, in both cases

there is a
competing hypothesis.
Regarding
the effect of

the modern rule
,

the alternative is that
fishers

lie about the use of the
konou

in closed weeks
.
This scenario will cause the estimate of



to



7

The arrondissement is an administrative unit between the village level and the commune level. Each of the four
arrondissements included in our sample com
prises three samples villages.

be negative
,

even
when

there is no

actual compliance to the
modern
rule
.
Regarding the
impact of Voodoo, there may
also
be

unobserved
time
-
invariant
factors

that

are correlated
with
both
adherence to
Voodoo

and the use of
fine
mesh

fishing
nets
. Such factors
would bias
the estimate
for


.

For instance, t
he
coefficient
estimate
may
be driven

by
certain
environmental factors
specific to villages with a high share of Voodoo followers
.
B
ecause of
the small number and proximity of the villages in
our

sample
, we do not
include village fixed
effe
cts in the above specification (
only arrondissement fixed e
ffects). This specification
therefore does not allow us to control for the impact of village
-
level variables.


To deal with the first of these caveats

(untruthful answerin
g)
, we examine the fluctuation of
fishing revenue
for shrimp
across weeks.
As the
konou

is
the most productive instrument

used for shrimp
fishing
, any periodical
variation in its use should be reflected in fishing
revenue

for shrimp
8
.
Thus, i
f

fishers
respect the modern rule and refrain from using the
konou

when the lakes are closed
, we sh
ould find that
shrimp
fishing revenue is lower in closed
weeks.
Moreover,
the resulting decline
of
fishing intensity should allow the fishery stock to
increase in the
closed period. F
ishing revenue should
then
be highest in the first week of re
-
opening, and gradually decline over the following open weeks as intensive
fishing with
the
konou

again
decreases

the fishery stock.



To analy
s
e the fluctuation of fishing
revenue across weeks, w
e estimate two regression
specifications. In the first specification we
examine weeks in which the lakes are closed
,
taking the open period as the baseline category. In the second specification, we look
at weeks



8

Total fishing revenue on the other hand is generated by a larger variety of fishing gear and will be considerably
less affected by changes in the use of the
konou
.

in which the lakes ar
e open
, taking the closed period as the baseline category. These two
specifications can be written as follows:









































(2)




















































(3)

where



denotes
shrimp
fishing revenue for fisherman
i

in week
t
;





and





are
dummy variables that take value 1 if the lake is closed, respectively open in week
t
for the w
-
th
consecutive week;




denotes a time trend;





is a vector consisting of

the following set of
control variables

for individual
i

in week
t
:

the logarithms of the number of fishing days, the
number of persons fishing and the total value of fishing instruments used
, and a dummy
variable indicating the use of shrimp pots
;



captures individual fixed effects
;



denotes a
classical random
error term clustered at the individual level.
The time trend is again included
to capture
variation due to
the
growth cycle

of shrimp
. The control variables in




reflect the
production function of the fishing activity
, capturing the input of time invest
ed, labo
u
r and
capital
.

We specifically control for the use of shrimp pots because this
fishing instrument is
one of the main generators of shrimp fishing revenue.


Significantly negative estimates for


and



would
suggest that fishermen respect the
modern rule

and use the
konou

less when the lakes are closed
. Significant
ly

positive estimat
es

for


,


,



and



that decrease in magnitude would
indicate
that
fishing revenue

is
higher when the lake
s

are

open and especially so in the beginning of
the open period,
provid
ing

additional evidence that fis
hermen respect the modern rule.


To deal with the second caveat (unobserved time
-
invariant factors correlated with Voodoo),
we make use of a larger dataset taken from the 2006 fishery census
, which

con
tains
information on

religio
us affiliation

and the use of fishing gear for 18,683 fishermen across
121 villages in all 9 communes surrounding Lake Nokoué and Lake Porto Novo.
This
extended dataset allows

us to examine the relation
between
Voodoo

and the use of the
konou

while controlling for
village
-
level variables through the inclusion of village fixed effects
.

We

estimate the following regression specification:






























(4)




is an
indicator variable taking value 1 if individual

i
reports

to

use the
konou

for
fishing, and 0 otherwise;



is an indicator variable taking value 1 if individual

i
reports
his religion to be
Voodoo
, and 0 otherwise;




i
s a set of control var
iables consisting of the
logarithm of age and the level of education reported by individual

i
;



indicates fixed effects
at the village level
;



is a classical random error term.
If the estimate of



is statistically
insignificant when village fi
xed effects
are included,
it

wou
ld suggest that any relation

between
adhering to
Voodoo and the use of the
konou

is mainly driven by village
-
level
variables.


It
can
also
be argued that the individual choice of religion is correlated with several
(unobserved) individual characteristics. We therefore replace the individual measure of
Voodoo used so far with a measure of the proportion of Voodoo followers in the village
.

We
estimate the following specification
:





































(5)

w
here








measures the number of Voodoo followers relative to the total number of
individuals in the village
, and



again
denotes
arrondissement fixed effects (instead of
village fixed effects)
.
If the
estimate for





is

insignificant
, it would
suggest that
any negative
relation

between
individual

adherence to

Voodoo and the use of the
konou

is mainly driven
by unobserved individual characteristics.


Yet, i
f

we find that

the proportion of Voodoo followers in the village significantly affect
s

the
use of the
konou
, an

additional

question arises
.
Is compliance to the Voodoo rule driven by
community mechanisms such as herding behavio
u
r
, or do
individual

motivations

such as
individual

beliefs
also matter?

To
a
d
dress

this question, we estimate a
last

specification in
which we include both the individual measure of Voodoo and the proportion of Voodoo
followers in the village. The specification can be written as follows
:








































(6)

I
f

the estimate for




is

significantly negative while
the estimate for




is insignificant, it
suggests that
the negative effect of

Voodoo
on the use of the
konou

is

driven mainly

by

community mechanisms
,

rather than
by individual motivations
.


4.
DATA AND SUMMARY STATISTICS

We make

use of data from a household survey implemented
in 2009 among fishing
communities at Lake Nokoué and Lake Porto

Novo
.

The survey
collected detailed
information
for

the members of
360

households
. The households
w
ere selected by taking a

stratified
random sample from
the

2006 fishery census

in

twelve

villages
, which

are located
in

four different
arrondissements

that
form part of

two

communes
.

The two communes


So
-
Ava
and

Aguesgues



are
located
in the vicinity

of
L
ake Nokoué and
L
ake

Porto Novo
, as is
illustrated

in Figure 1
.


The data used for the empirical analysis
cover

the household members whose main
occupation is fishing
; it
include
s

weekly information on fishing activit
ies

and information on
individual and household characteristics for
255

individuals.
The
fishermen were visited bi
-
weekly during a period of

14 weeks, resulting in high
-
frequency data on fishing activities for
14 consecutive weeks.
For the estimation
of equations
(
4) to (6)

w
e make use of data from the
2006 fishery census,
which includes
one observation
per individual
with information on
religious affiliation and the use of fishing gear

for

18,683
individuals involved in the fishery
sector. These individuals are distributed across 121 villages in all 8

communes surrounding
Lake Nokoué and Lake Porto Novo.

For the empirical analysis we will

use data

on fishermen

only; the
sample of fishermen

includes 12187
individuals

distributed across 121 villages in
all 8 communes
9
.


R
eligious affiliation

Table
1

presents the
distribution of religious affiliation

across the two lakes,
both
for the
survey
sample of fishermen and the
(complete)
2006 fishery census
.
The overall distribution
is quite similar
, though

Panel B shows that t
here are some differences with respect to the
ranking within the lakes. At
L
ake Nokoué the three main religions are Voodoo, Ca
tholicism
and
the Celestial Church of Christ

(
a prophetic church founded in Benin)
. At
L
ake Porto
Novo, the main religions are Catholicism, Protestantism and other
forms of Christianity

(in
the fishery census) or Islam (in the survey sample), while
Voodoo is only of minor
importance. The large heterogeneity in religious affiliation can be
partly

explained by the
historical migration patterns of different ethnic groups
, as well as
the

religious pluralism and

high tolerance towards individual religious

choice in Benin

(
Barbier & Dorier
-
Apprill, 2002
).

The
exceptional
religious
heterogeneity

is furt
her borne out by the fact that
i
n
the household
survey

we find within
-
household variation in reported religion for about 31

percent

of the
households.




9

The sample of fi
shers includes full
-
time fishermen, part
-
time fishermen

who are also involved in farming or
keeping cattle

and seasonal fishermen
.

The fishing activity in this region is dominated by men: less than 3
percent

of the indivi
duals in the dataset are women.


Use of
the
konou

Table 2 present
s

information on the use of the
konou

in the survey sample and the census
sample
.
In the census sample, 16.6 % of the individuals use the
konou

for fishing. In the
survey sample

the
share of
konou

users is considerably larger: about 40 % of the fishermen
used the
konou

at least
once during the sample period
.

Figure 2
shows
the evolution of the
use of the
konou

across weeks: the bars indicate the share of
konou

fishers

that used the
konou

for fishin
g
per

week
10
. Although

the use of the
konou

is lower

when the lakes are
closed,
the drop i
s

modest
.

The Wilcoxon
-
Mann
-
Whitney test for the difference between the
average use of the
konou

among

konou

fishers in open and closed weeks rejects the null of
equal means at the 1 % level (with a test
-
statistic of 3.584 and p
-
value of 0.0003).

Figure 3 depicts the fluctuation of fishing revenue for shrimp across open and closed weeks,
showing that
average rev
enue drops when the lakes are closed.

For the second
time

of
closing
, however, fishing revenue starts to rise again before the lakes are re
-
opened.
Two
alternative explanations can be offered for this pattern.

First, fishermen

may be shirking
more
when the benefits
are higher
, that is
,

when
the quantity of fish

has
increased

after

the first
week of closure
.
Alternatively,
fishermen may
start to reap the benefits

of the
konou

ban
already in the second week of closure in the form of

an increase
in

fis
hing

revenue
,
an effect
that may be stronger at the end of the fishing season as the fishery stocks decrease
.


Control variables




10

We identify a
konou

fisher as a fisherman

that has used the
konou

at least once during the survey period. We
do not
use
ownership of the
konou

because

fishers occasionally borrow or rent the
konou

from others.
Ownership therefore is not equivalent to the actual use of the
konou
.

Table 3

presents summary statistics of
the
characteristics of fishermen

and

weekly
fishing
activities
.

It is worth noting that the literacy rate
of 14.3

percent
lies far below the
estimated
national
average of
55.2
percent
for male adults
11
.
Another noteworthy feature

is the
strong
dependency
of
income

on fishing:

91

percent
of yearly
household
income
consists of income
from the fishery sector
.


5.
RESULTS

The Breusch
-
Pagan/
Cook
-
Weisberg test for heteroskedasticity

indicates conditional
heteroskedasticity of the error terms for all specifications (Chi
2
(1)=
292.04
***). The
Wooldridge (2002) test for seri
al correlation in panel
-
data models indicates
that there is serial
correlation in the

error terms (F(1,230)=404.403***)
for

the

specifications
related to

the
household survey sample
.
We
use standard errors clustered at the individual level to correct
for
heteroskedasticity and serial correlation of the error terms

when estimating equations 1


4

(household survey sample)
, a
nd
use simple
heteroskedasticity
-
robust standard errors
when
estimating equations 5 and 6 (
fishery

census sample
)
.


Compliance with th
e
regulations

implies that



and



should be negative for our
first
specification
(equation 1)
:

the
konou

should be used less in w
eeks in which the lake
s

are
closed
, and should be used less by
V
oodoo
followers
.

Table
4

presents
the
OLS estimation
r
esults
for
equation

(1)
.

The dependent variable



equals

1 if the
konou

is
among

the
self
-
reported fishing instruments used in week t. If we redefine the dependent variable such
that it takes value 1 only if the
konou

is

the most important
instrument used in week t, the
results are qualitatively the same
.




11

2010 estimate of the Unesco Institute for Statistics (UIS).

We find that

1

is negative and statistica
lly significant in all columns. This result confirms the
first hypothesis:

the use of the
konou

is on average lower in weeks in which the lakes are
closed

(
compared to open weeks
)
.
T
he estimates of



are
also
negative and statistically
significant in all colum
ns,
indicating
a negative
relation

between
Voodoo

adherence
and the
use of the
konou
.

The results in column (6) suggest that the probability of using the
konou

is
on average
13.1

percent lower for Voodoo fishermen, compared to fishermen who adhere to
other religions

(all else equal).
With respect to the modern rule, the results suggest that

the
probability of using the
konou

in closed weeks is on average 2
0

percent
lower
, compared to
open weeks (all else equal)
.
The insignificance of the coefficient for the literacy interaction
term implies that compliance does not differ between literate an
d illiterate fishermen.


With respect to
the specifications for
fishing

revenue across weeks,
our hypothese
s state that



and

2

are negative,
while


,


,

5

and



are positive

and decreasing in size
. The
Hausman specification

test indicates that the fixed e
ffects model is most appropriate to
estimate these specifications.
The
Wooldridge test for serial correlation in panel
-
data models

indicates that there is
serial correlation in the error terms, so we cluster standard errors at the
individual level.
Table
5

presents the f
ixed effects estimates

for specification (2)
;
T
able 6
presents the fixed effects estimates for specification (3).


In
all columns

of Table
5

the estimates of



is

negative and statistica
lly significant: fishing

revenue
for shrimp
is
on average
lower in
the first week of closing

compared to

open weeks.
This negative effect, however, vanishes in the second week of closing.

The results presented
in
Table 6
support

the hypotheses as well.
In
all

column
s

the
estimates
of



and




are positive and statistica
l
ly significant
, suggesting that
fishing

reve
nue

for
shrimp

is
on average
higher in
the first and second open
weeks

comp
ared to the closed
period.
The positive impact of opening the lakes is smaller in the second week of opening
compared to the first week, and vanishes in the third (and fourth week).


Regarding
the
specifications for the
2006 fishery census

sample
, our hypotheses
state

that



and




are significantly negative. T
able 7

presents the OLS estimation results for
equation

(4) in columns (1) to (3). Column (4) presents the
OLS estimates for equation (5) and column
(6) shows the OLS estimates for equat
ion (5).


A comparison of Table
4

and Table
7

reveals

that the estimation results for
the
2006
fishery
census
sample
are

in line with the results for
the 2009

household survey

sample. The
estimates of



are negative and statistically
significant in
all

columns.

When c
ontrolling for
village
fixed effects
in column (2),

the coefficient for Voodoo

remains significantly negative
.
Controlling for arrondissement fixed effects in column (3) does not change the
significance

of the coefficient for Voodoo

(and ba
rely changes
its

magnitude)
.
These estimates rule out
the possibility
that the
negative coefficient
of

Voodoo is

driven by village level variables.
In
column (4), where the individual measure of Voodoo
is replaced
by the share of Voodoo
followers in the village, the estimate of




is significantly negative as well.
This finding
suggests that the negative relationship between Voodoo and the use of the
konou

is not
reflecting the impact of (unobserved) individual c
haracteristics.

Finally, in column (5)
, where

both the individual measure and the village share of Voodoo are included
, the

coefficient
s

for
both
measures
are

significantly
negative
. These results suggest

that the

negative effect of
adherence to Voodoo on
the use of the
konou

is
driven by both
community

mechanisms

and
individual motivations.


6.

ROBUSTNESS CHECKS

The first robustness check

concerns the results for the use of the
konou

across weeks.
W
e

test
whether the results obtained
with the

OLS model hold under alternative
regression

models.
Since it is intuitive to interpret

the dependent variable as a probability,
we start by estimating
equation (1) using a probit model

(results not reported)
.
The average marginal effects of the
probit
estimation results are presented in Table 8.

A comparison of

Table
8
and

Table 4

shows that the results of the OLS and probit model
are
qualitatively the same.
The coefficient for the closing of the lake is negative and
statistically
significant

(p
-
value of 0.05)
, suggesting
that the

probability of using the
konou

is on average
lower
when the lakes are

closed
, compared to open weeks.
The

coefficient for
Voodoo

is
also
negative and statistically significant

in all columns

(p
-
value of 0.07)
,
suggesting

that Voodoo
fishermen on average have a lower probability of using the
konou

compared
to
other
fishermen.

The
average marginal effects
for
the
Voodoo and
closed variable
s

have
roughly
the same
magnitude
as the corresponding OLS estimates
.

The re
sults for the closed variable also hold when

we estimate specification (1) using a
conditional fixed effects logit model

(results not reported).



As a second robustness check, w
e perform

a falsification test for the results

for
fishing
revenue across
weeks
. More specifically, we
examine

whether the coefficients for the week
dummies are picking up the impact of some variable
other
than the opening and closing of the
lakes. To do so, we estimate specifications (2) and (3)
using

data for Lake Ahémé instea
d of
Lake Nokoué and Lake Porto Novo.
Lake Ahémé is another important water body

in
s
outhern

Benin
and offers an in
teresting case for comparison.
The lake

is
rather

similar to
Lake Nokoué and Lake Porto Novo
with respect to the

nature of fish
ing activities
, but


for
our purpose


t
he
key difference is
the
fact that
the
konou

is not used at Lake Ahémé
. Hence,
the modern rule
does not exist
here
.
If

we repeat

the estimation of
equations

(2) and (3)
using
data for

Lake Ahémé
, we should find
no

particular pattern
for
the coefficients



,


,


,


,



and


. If we do find such a pattern, it may indicate

that
our results for
L
ake Nokoué
and
Lake
Porto Novo
are not
driven by
compliance to the
modern rule
, but
rather
by

a natural
phenomenon that characterizes the fishing activity
at
the
se

lakes
.



Figure 5 shows the evolution of average weekly

fishing revenue for shrimp at L
ake Ahémé
and gives a first indication that there are no systematic fluctuations of fishing revenue

at this
lake. Average revenue does not drop
strongly
in weeks in which Lake Nokoué and Lake
Porto Novo are closed, nor does it rise again in weeks in which th
ese lakes are reopened.
Table 9

presents the fixed effects estimation results for
equation (2) an
d Table 10 presents t
he
fixed effects estimation results for
equation

(3) for the
Lake
Ahémé sample
.

The coefficient
for the first week of closing

(




is negative and statistically significant, but none of the
coefficients for the open week dummies are
significantly different from zero. This indicates
that the decrease in fishing revenue in the closed week does not bring about a subsequent rise
in fishing revenue or in fishing yield. Moreover, the significance of the coefficient for the
first week of clo
sing does not arise until we add the time trend to the specification. These
observations lead us to believe that the negative coefficient for the first week of closing
captures the impact of some event coinciding with the closed weeks at lake Nokoué and La
ke
Porto Novo. One such event could be for instance a large Voodoo festival, which would
induce all Voodoo fishermen


or
90 % of the fishermen at lake Ahémé


to stop or
limit

their fishing activity during the event.

Amoussou (2004) for instance mentions

that fishing
is
prohibited

for

one week
in

all fishing villages at lake Ahémé during the annual festivities in
honour of the Voodoo spirit Avlékété.


7.

DISCUSSION AND CONCLUSION

The coastal fisheries of Benin provide a textbook example of
a tragedy of th
e commons.
Intensive fishing by individuals pursuing their self
-
interest has compromised the
sustainability of the fishery stock, which has dramatically declined in the last decades
(Amoussou, 2004; FAO, 2008; Gnohossou, 2006; USAID, 2007). This overexploi
tation is a
direct
consequence of socio
-
economic

changes
, such as the
steady growth

of the number of
fishers and a rapidly rising
demand for
fishery products.
A
nother important cause of the
overfishing problem is the erosion of the traditional Voodoo
governance system
, which is
itself a consequence of the profound socio
-
economic changes in south Benin
(Bourgoignie,
1972; Dangbégnon, 2000; Pliya, 1980). Attempts by the state to
fill the institutional vacuum
by
creating alternative governance in
stitution
s were unsuccessful,
failing
to
effectively
regulate
fishing activities

(Dangbégnon, 2000).

From the 1960s onwards, the institutional weakness and rising value of fishery products
resulted in the introduction and proliferation of high
-
yielding but damaging

fishing
instruments, which exacerbated the problem of overfishing (
Amoussou, 2004; Dangbégnon,
2000; FAO, 2008; Gnohoussou, 2006; USAID, 2007
). The
konou

(also known as

medokpo
konou
)

is one such instrument and form
ed

the object of this study
. In response
to the
problems of overfishing and resource degradation, local fishing committ
ees were formed in
recent years. These committees
created

a new rule for the
use of the
konou
, allowing its use
for four consecutive
weeks, then banning

it for the following two
weeks.

The
traditional
Voodoo institution

also stipulates rules about the use of fishing gear and so
prohibit
s

the use of fine mesh
fishing nets, a rule that applies to the
konou

(Dangbégnon,
2000; Clédjo, 2006; Pliya, 1980).


For the case of the coastal
lakes of Benin, we have shown that
the
Voodoo
-
based traditional
institutions survived
a rapidly changing environment
at least to some extent
and still play a
role in regulating
fishing activities
. One aspect that may have contributed to its resilience is
the strong integration of Voodoo in Beninese culture. The deep roots of Voodoo
were
recently

materialized by its enlistment in the
Beninese
constitution as an official religion and
the existence o
f a
national
public
festival

to celebrate Voodoo.


Although we have found a strong correlation between refraining from using of the
konou

and
adherence to Voodoo



and this correlation is robust to the use of different samples, the
inclusion of village fi
xed effects, and the use of village
-
level rather than individual
-
level
Voodoo adherence


there are competing explanations. For instance, the observed correlation
could

be caused by
conservatism

among Voodoo followers
;
rather than the impact of the rule
p
er se, we
may
observe the ef
fect of aversion towards change

in the form of new fishing gear
and

new monotheistic religions
.

Further research should attempt to investigate this
possibility.


We have found evidence for a statistically significant impact of t
he modern rule on the use of
the
konou
.
However, quantitatively

the impact is small. One reason for the observed
compliance is that
konou

fishers have a short
-
term
incentive to respect this rule, as
c
ollectively
refraining

from
using

the
konou

for some weeks raises fishing
yield
. The reason
for the limited quantitative effect

is likely to originate from
the free rider problem. Little
monitoring and the absence of effective sanctions create strong incentives to free ride on the
compliance of oth
er
konou

fishers and maximally reap the benefits from the increase in
fishing yield.


It is not straightforward to

interpret

these results in terms of ecological effectiveness of the
modern rule.
T
wo

weeks of closing
may not be

sufficient to allow the fish
ery stock to
reproduc
e,

but may

merely
resul
t

in larger quantities of fish and shrimp moving into the
lakes, as many of the indigenous species are migratory species (see Gnohoussou, 2006 and
Pliya, 1980). When these additional quantities are harvested in the open weeks before they
can reproduce, they

only cause local and short
-
term increases in the stock of fish. The
modern rule could then prove ineffective
in promoting

the regeneration of the fishery stocks

and support
ing

the sustainability of the fishery resource. Future research
could

investigate t
he
ecological impact of the existing rules
and

explore
the possible impact of alternative rules
,
such as

spatially
regulating

the use of the
konou

in the form of protected areas.




REFERENCES


Agrawal, A. (2001).
Common Property Institutions and Sustainable Governance of
Resources.
World Development
29 (10):

1649
-
1672.


Allan, J.D., Abell, R., Hogan, Z., Revenga, C., Taylor, B.W., Welcomme, R.L. &
Winemiller, K. (2005).
Overfishing of inland waters.
Bioscience

55(12):

1041
-
1051.

Amoussou, E. (2004). Systèmes tradition
nels de gestion durable du lac Ahémé au Bénin. In
Développement durable : leçons et perspectives

(pp. 263
-
270)
.
Acte de Colloque AUF
Ouagadougou, Benin.

Baland, J. M. & Platteau, J. P. (1996).
Halt
ing degrad
ation of natural resources
: is there a
role for rural communities?
Oxford: Clarendon Press.

Barua, I. (2009). Conservation and Management of Community and Natural Resources: A
Case Study from North East India.
Studies of Tribes and Tribals

7(1):

39
-
46.

Berkes, F. (Ed.) (1989). Common Property Resources. Ecology and Community
-
Based
Sustainable Development. London: Belhaven Press.

Berkes, F., Feeny, D., McCay, B.J. & Acheson, J.M. (1989). The Benefits of the Commons.
Nature
, 340:

91
-
93.

Berkes, F.,

Colding, J. & Folk, C. (2000). Rediscovery of traditional ecological knowledge as
adaptive management.
Ecological Applications
, 10(5):

1251
-
1262.

B
ourgoignie
, G.E. (1972) Les Hommes de l'Eau. Ethno
-
Écologie du Dahomey Lacustre.
Paris
:

Éditions
Universitaires
.

Bromley, D. (Ed.) (1992). Making the commons work: Theory, practice and policy.
San
Francisco CA: Institute for Contemporary Studies Press.

Clédjo F.G.A.P., (2006) :
La gestion locale de l’environnement dans les cités du lac Nokoué
au Bénin

méridional
. Thèse de Doctorat (Thèse Unique), Ecole Doctoral Pluridisciplinaire
FLASH DGAT UAC 321 p.


Cox, M., Arnold, G., & Villamayor Tomás, S. (2010).

A review of design principles for
community
-
based natural resource management
.


Ecology and Society

15
(4):

38.

Cummings, W. C. (1961).
Maturation and Spawning of the Pink Shrimp, Penaeus duorarum
Burkenroad
.
Transactions of the American Fisheries
Society

90
(
4):

462
-
468.


Dangbégnon, C. (2000).
Governing local commons: what can be learned from the failures of
lake Aheme’s institutions in Benin?

Paper presented at the eighth biennial conference of the
International Association for the Study of Common Property. May 31st
-

June 4
th

2000,
Bloomington, Indiana, USA.


Dietz, T., Ostrom, E. & Stern, P.C. (2003). The Struggle to Govern the Commons.
Scie
nce

302:
1907
-
1912.

Dorm
-
Adzobu,

C.,
Ampadu
-
Agyei
, O.

&
Veil
, P. G.

(1991).
Religious beliefs and
environmental protection: the Malshegu sacred grove in northern Ghana. Washington. DC,
USA: World Resources Institute and Nairobi. Kenya: Acts Press. African Centre for
Technology Studies.


Eneji C. V. O., Ntamu G. U., Unwanade C.
C., Godwin A. B, Bassey J. E., Willaims J. J &
Ignatius
, J.

(2012). Traditional African Religion in Natural Resources Conservation and
Management in Cross River State, Nigeria
. Environment and Natural Resources Research
2(
4): 45
-
53.


Food and Agriculture Organization (2012).
The state of world fisheries and aquaculture.

Rome: FAO Fisheries and Aquaculture Department.


Feeny, D., Berkes, F., McCay, B.J. & Acheson, J.M. (1990). The Tragedy of the Commons:
Twenty
-
Two Yea
rs Later. Human Ec
ology, 18 (1):

1
-
19.


Feeny, D., Hanna, S. & McEvoy, A.F. (1996). Questioning the Assumptions of the “Tragedy
of the Commons” Model of Fisheries.
Land Economics
, 72 (2):

187
-
205.


Food and Agriculture Organization (FAO) (2008).
Vue générale du secteur des
pêches
national, la republique du Bénin
.
Rome

: Fishery and Aquaculture Country Profiles.


Gibson, C.C., McKean, M.A. & Ostrom, E. (Eds.) (2000).
People and Forests: Communities,
Institutions, and Governance
.
Cambridge, MA: The MIT Press.

Gnohossou, P. M. (2006).
La faune benthique d’une lagune ouest africaine (le lac Nokoué au
Benin), diversité, abondance, variations temporelles et spatiales, place dans la chaine
tropique
. PhD Dissertation.
France:

Institut Nati
onale Polytechnique de Toulo
use
.


Hardin, G. (1968).
The Tragedy of the Commons.
Science

162:

1243

1248.


Kajembe,
G.C.,
Luoga,
E. J.,
Kijazi
M. S.
& Mwaipopo
C. S. (2003).
The role of traditional
institutions in the conservation of forest resou
rces in East Usambara, Tanzania.

International
Journal of Sustainable Development &
World Ecology
10(
2
):

101
-
107
.


de Kimpe, P. (1968).
Les facteurs de production piscicole des lagunes de l’Est Dahomey et
leur évolution récente
.
Côte d’Ivoire:
Centre Techn
ique Forestier Tropical, Bouaké.


Lalèyè, P.A. (2000).
Acadja

fisheries enhancement systems in Benin: their productivity and
environmental impacts. In Abban, E.K.,
Casal, C.M.V., Falks, T.M. & Pullin
, R.S.V. (Eds.),

Biodiversity and sustainable use of fish in the coastal zone

(pp. 51
-
52). ICLARM.

National Research Council (NRC) (2002).

The Drama of the Commons.

Ostrom,
E.,
Dietz,
T.,
Dolšak,
N.,
Stern,
P.C.,
Stovich,
S. &
Weber,
E.U.
(Eds.), Washington, DC: National
Academy Press.

Niyonkuru, C. & Lalèyè, P.A. (2010). The impac
t of
acadja

fisheries on fish assemblages in
Lake Nokoué, Benin, West Africa.
Knowledge and Management of Aquatic Ecosystems

339:

05.

Accessed
April 25, 2012.
doi: 10.1051/kmae/2010033
.


Maarleveld, M. & Dangbégnon, C. (1999). Managing natural resources: A

social learning
perspective.

Agriculture and Human Values
16
:

267

280.


Niyamweru, C. & Sheridan, M. (Eds.) (2008). African Sacred Ecologies.
Journal for the
Study of Religion, Nature and Culture

2(3): 285
-
407.



Ostrom, E. (1990).
Governing the Commons:
The Evolution of Institutions for Collective
Action
. Cambridge NY: Cambridge University Press.


Ostrom, E. (1999). Coping with tragedies of the commons.
Annual Review of Political
Science
2:

493
-
535.


Partenariat National de l’Eau du Bénin (PNE) (2010).
Défis et orientations stratégiques pour
la gestion des conflits liés aux engins et techniques de pêche sur le lac Ahémé et le complexe
lac nokoué
-
lake Porto Novo au Bénin
.
Benin, n
ote technique.

Pinkerton, E. & Weinstein, M. (1995).
Fisheries That Work: Su
stainability Through
Community
-
Based Management
(Report).
Vancouver, British Columbia

: David Suzuki
Foundation.


Pliya J. (1980)
.

La pêche dans le Sud
-
Ouest du Bénin. Etude de géographie appliquée sur la
partie

continentale et maritime
.
Paris: AGECOOP.


Pliya, J. (1989).
Migrations historiques et peuplement dans les régions lagunaires du Bénin
méridional. In :
Antheaume B.
,
Blanc
-
Pamard C.,

Chaléard J.L.,
Dubresson A.
,
Lassailly
-
Jacob V.
,
Marchal J
-
Y.
,
Pillet
-
Schwartz A.M.
,
Pourtier R.
,
Raison J
-
P.
,
Sevin

O.

(Eds.),
Pinton F. (coord.)

Tropiques : lieux et liens : florilège offert à Paul Pelissier et Gilles Sautter

(pp. 525
-
531).
Paris: ORSTOM.


Rusinga, O. & Maposa, R. (2000).
Traditional religion and natural resources: A reflection on
the significance of
indigenous knowledge systems on the utilisation of natural resources
among the Ndau People in South
-
eastern Zimbabwe
.

Journal of Ecology and the Natural
Environment

2(9):

201
-
206.


Sharma, S., Rikhari H. C. &
Palni
, L. M. S.

(1
999).
Conservation of Natural Resources
Through Religion: A Case Study from Central Himalaya.
Society & Natural Resources: An
International Journal

12(6):

599
-
612
.


Stoop, N., Verpoorten, M., Houssa, R.
(
mimeo
)
.
To fish or not to fish? Resource degradation
and
income diversification in Benin.


Tall, E. K. (1995a).
Dynamique des cultes voduns et du Christianisme céleste au Sud
-
Bénin.
Cahiers des Sciences Humaines

31
(4): 797
-
823.


Tall, E. K. (1995b). De la démocratie et des cultes voduns au Bénin.
Cahiers
d'études
africaines

35(137): 195
-
208.


United States Agency for International Development (USAID) (2007).
118/119

Biodiversity
and tropical forest assessment for Benin.


Wade, R. (1994).
Village Republics: Economic Conditions for Collective Action in South
India.
San Francisco CA: ICS Press.


Williams, A.B. (1984).
Shrimps, lobsters and crabs of the Atlantic coast of the eastern United
States, Maine to Florida.
Washington, D.C

:
Smithsonian Ins
titution Press.











LIST OF FIGURES AND TABLES


Figure 1: Map of the survey area



Lake Ahémé

Lake Nokoué

Lake Porto Novo




Cotonou

Porto Novo

Kpomasse

So
-
Ava

Aguesgues

Figure
2
: Use of the
konou

across open and closed weeks

for
konou

fishers


Figure 3
:
Average weekly fishing revenue for shrimp across open and closed weeks


0
10
20
30
40
50
60
% of konou fishers using the konou
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Week
Light-gray bars indicate weeks in which the lakes are closed
(% of individuals)
Use of the konou across weeks for konou fishers




0
2000
4000
6000
8000
10000
12000
Average weekly fishing revenue (FCFA)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Week
Gray areas indicate weeks in which the lakes are closed
Average weekly fishing revenue for shrimp across weeks





Figure 4
:
Average weekly fishing revenue for shrimp at Lake Ahémé: counterfactual












0
2000
4000
6000
8000
10000
12000
Average weekly fishing revenue (FCFA)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Week
Gray areas indicate weeks in which Lake Nokoué and Lake Porto Novo are closed
Lake Ahémé
Average weekly fishing revenue for shrimp across weeks






Table 1: Distribution of religious affiliation across lakes: survey sample and fishery census

Panel A: Survey sample distribution of religious affiliation across lakes (% of individuals)

Religion

Lake Nokoué

Lake Porto Novo

Overall

N

Catholicism

27.2

44.7

37.5

95

Protestantism

11.7

25.3

19.8

50

Islam

0.0

16.0

9.5

24

Voodoo

27.2

2.7

12.6

32

Christianisme Céleste

22.3

4.7

11.9

30

None

6.8

0.0

2.8

7

Other

4.9

6.7

5.9

15

Total

100

100

100

253

Panel B: Fishery census
distribution of religious affiliation across lakes (% of individuals)

Religion

Lake Nokoué

Lake Porto Novo

Overall

N

Catholicism

25.1

33.0

27.0

4,995

Protestantism

12.0

21.0

14.2

2,628

Islam

3.2

9.7

4.8

885

Voodoo

21.7

5.8

17.8

3,294

Christianisme
Céleste

18.4

8.9

16.1

2,981

Other Christianities

10.3

18.2

12.3

2,268

None

9.3

3.3

7.9

1,457

Total

100

100

100.0

18,508


Table 2:
Share of
konou

users (fishermen who used the
konou

at least once during 14 weeks)

Week

Freq.

N

% N

Survey sample

120

255

47.1

Census sample

1,280

7,731

16.6


Table 3: Characteristics of fishermen and weekly fishing activities in household survey sample

Variable

Mean

St. Dev.

Max

N

Age

42.8

13.2

90.0

242

Literacy (% of fishermen)

14.3





252

Total value of assets

29
,026

43,123

427,000

255

Yearly income

1
,
437
,077

1,205,328

7,850,000

255

Yearly income from fishery sector

1
,
310
,951

1,110,275

7,850,000

255

Number of fishing days

4.0

2.3

7.0

3,570

Number of persons fishing

2.2

1.9

23.0

2,828

Total value of fishing
gear used

126,321

259,394

3,090,000

3,569

Notes: Assets consist of the number of motorbikes, bicycles, radios and mobile phones owned by the
household. Earnings are expressed in FCFA. One euro is approximately 656.66 FCFA.


Table 4

OLS estimation results: Use of the
konou

across weeks

Dependent variable: Use of the
konou

Variables

(1)

(2)

(3)

(4)

(5)

(6)

(7)

















Voodoo

-
0.127**

-
0.127**

-
0.097*

-
0.101*

-
0.099*

-
0.131*

-
0.131*



(0.051)

(0.051)

(0.053)

(0.054)

(0.054)

(0.071)

(0.071)

Closed

-
0.051***

-
0.051***

-
0.052***

-
0.051***

-
0.050***

-
0.049***

-
0.200**



(0.013)

(0.013)

(0.014)

(0.014)

(0.014)

(0.014)

(0.101)

Time



-
0.000

-
0.000

-
0.001

-
0.001

-
0.000

-
0.000





(0.003)

(0.003)

(0.003)

(0.003)

(0.003)

(0.003)

Log(Age)





-
0.309***

-
0.317***

-
0.302***

-
0.191***

-
0.191***







(0.067)

(0.067)

(0.070)

(0.062)

(0.062)

Literacy







-
0.058

-
0.054

-
0.026

-
0.051









(0.056)

(0.056)

(0.052)

(0.060)

Log(Assets)









0.024*

0.023*

0.023*











(0.013)

(0.012)

(0.012)

Literacy*Closed













0.082















(0.052)

Arrondissement

No

No

No

No

No

Yes

Yes

















Constant

0.278***

0.280***

1.435***

1.578***

1.452***

1.012***

1.059***



(0.024)

(0.030)

(0.253)

(0.264)

(0.287)

(0.265)

(0.269)

















Observations

2,976

2,976

2,807

2,768

2,768

2,768

2,768

R
-
squared

0.014

0.014

0.065

0.073

0.080

0.182

0.183

Notes: Coefficients are reported with individually clustered standard errors in parentheses. ***, ** and *
denote
significance at the 1, 5 and 10 % levels respectively. The acadja is excluded from the regression
sample. Arrondissement fixed effects are included in column (6)

and (7)
. Assets are expressed in 1000 CFA.
One euro equals about 656 CFA.












Table 5

Fixed effects estimation results: Fishing revenue across weeks
-

Closing of lakes

Dependent variable: Logarithm of fishing revenue for shrimp

Variables

(1)

(2)

(3)

(4)

(5)











Closed first week

-
0.715***

-
0.757***

-
0.567***

-
0.577***

-
0.555***


(0.184)

(0.182)

(0.159)

(0.201)

(0.184)

Closed second week

-
0.351**

-
0.416**

-
0.241*

-
0.193

-
0.148


(0.162)

(0.162)

(0.143)

(0.183)

(0.168)

Time trend


-
0.169***

-
0.164***

-
0.237***

-
0.239***



(0.031)

(0.028)

(0.036)

(0.035)

Log(Fishing days)



2.633***

2.070***

1.755***




(0.182)

(0.521)

(0.498)

Log(Persons fishing)




0.412

0.460





(0.577)

(0.586)

Log(Value fishing gear)





0.395***






(0.066)

Shrimp pots





5.165***






(0.535)

Constant

3.783***

5.150***

1.513***

2.713**

-
1.361


(0.055)

(0.249)

(0.333)

(1.132)

(1.260)







Observations

3,048

3,048

3,048

2,309

2,309

R
-
squared

0.006

0.039

0.203

0.072

0.191

Number of individuals

255

255

255

247

247

Notes: Coefficients are reported with individually clustered standard
errors in parentheses. ***, ** and *
denote significance at the 1, 5 and 10 % levels respectively. The acadja is excluded from the regression
sample.















Table 6






Fixed effects estimation results: Fishing revenue across weeks
-

Opening of
lakes



Dependent variable: Logarithm of fishing revenue for shrimp



Variables

(1)

(2)

(3)

(4)

(5)











Open first week

1.161***

1.249***

1.035***

1.116***

1.067***


(0.209)

(0.208)

(0.180)

(0.223)

(0.210)

Open second week

0.739***

0.996***

0.840***

0.917***

0.729***


(0.238)

(0.235)

(0.206)

(0.264)

(0.246)

Open third week

0.405*

0.576***

0.281

0.236

0.162


(0.222)

(0.218)

(0.195)

(0.246)

(0.238)

Open fourth week

0.236

0.297*

0.110

0.079

0.149


(0.166)

(0.165)

(0.147)

(0.186)

(0.180)

Time trend


-
0.176***

-
0.170***

-
0.242***

-
0.241***



(0.031)

(0.028)

(0.036)

(0.035)

Log(Fishing days)



2.633***

1.970***

1.665***




(0.180)

(0.527)

(0.503)

Log(Persons fishing)




0.427

0.496





(0.569)

(0.579)

Log(Value fishing gear)





0.379***






(0.066)

Shrimp pots





5.175***






(0.533)

Constant

3.199***

4.513***

1.074***

2.426**

-
1.500


(0.114)

(0.279)

(0.352)

(1.135)

(1.270)







Observations

3,048

3,048

3,048

2,309

2,309

R
-
squared

0.011

0.047

0.210

0.082

0.199

Number

of individuals

255

255

255

247

247

Notes: Coefficients are reported with individually clustered standard errors in parentheses. ***, ** and *
denote significance at the 1, 5 and 10 % levels respectively. The acadja is excluded from the regression
sample.













Table 7






OLS estimat
ion results: Use of the
konou

-

2006 fishery census

sample



Dependent variable: Use of the
konou


Variables

(1)

(2)

(3)

(4)

(5)







Voodoo

-
0.063***

-
0.076***

-
0.081***


-
0.073***


(0.010)

(0.011)

(0.011)


(0.011)

Share of Voodoo in village




-
0.185***

-
0.110**





(0.046)

(0.047)

Log(Age)

-
0.177***

-
0.096***

-
0.115***

-
0.128***

-
0.116***


(0.011)

(0.011)

(0.011)

(0.011)

(0.011)

Education

-
0.025***

0.002

-
0.002

-
0.001

-
0.003


(0.004)

(0.004)

(0.004)

(0.004)

(0.004)

Village

No

Yes

No

No

No







Arrondissement

No

No

Yes

Yes

Yes







Constant

0.832***

0.399***

0.483***

0.586***

0.539***


(0.042)

(0.047)

(0.044)

(0.050)

(0.051)







Observations

7,707

7,707

7,707

7,707

7,707

R
-
squared

0.039

0.340

0.178

0.175

0.179

Notes: Coefficients are reported with robust standard errors in parentheses. ***, ** and * denote significance
at the 1, 5 and 10 % levels respectively. The acadja is excluded from the regression sample. Village fixed
effects are
included in column (2). Arrondissement fixed effects are included in columns (3), (4) and (5).















Table 8






Probit estimation results: Use of the
konou

across weeks
-

Average marginal effects

Dependent variable: Use of the
konou






Variables

(1)
















Voodoo

-
0.128*






(0.071)





Closed

-
0.185*






(0.094)





Time

-
0.000






(0.003)





Log(Age)

-
0.183***






(0.055)





Literacy

-
0.048






(0.050)





Log(Assets)

0.024**






(0.012)





Arrondissement

Yes











Literacy*Closed

0.072






(0.049)











Observations

2,768









Average mar
ginal effects are reported with

individually clustered standard errors in parentheses. ***,
** and * denote significance at the 1, 5 and 10 % levels respectively. The acadja is excluded from the
regression sample.















Table 10











Fixed effects estimation results:
Fishing revenue across weeks at lake Ahémé

Dependent variable: Logarithm of fishing revenue for shrimp





Variables

(1)

(2)

(3)

(4)

(5)











Closed first week

-
0.250

-
0.428**

-
0.258*

-
0.299*

-
0.321**


(0.191)

(0.188)

(0.143)

(0.155)

(0.154)

Closed second week

0.357***

0.302**

0.170*

0.163

0.147


(0.130)

(0.131)

(0.101)

(0.103)

(0.099)

Time trend


-
0.178***

-
0.169***

-
0.171***

-
0.184***



(0.038)

(0.031)

(0.034)

(0.033)

Log(Fishing days)



4.084***

3.011***

2.899***




(0.158)

(0.834)

(0.787)

Log(Persons fishing)




2.790***

2.787***





(0.611)

(0.583)

Log(Value fishing gear)





0.150**






(0.068)

Shrimp pots





1.100






(0.693)

Constant

6.910***

8.376***

1.079***

0.932

-
0.226


(0.052)

(0.309)

(0.406)

(1.617)

(1.741)







Observations

1,652

1,652

1,652

1,518

1,518

R
-
squared

0.005

0.058

0.368

0.143

0.153

Number of individuals

118

118

118

118

118

Notes: Coefficients are reported with individually clustered standard errors in parentheses. ***, ** and *
denote
significance at the 1, 5 and 10 % levels respectively. Lake Nokoué and lake Porto Novo are excluded
from the regression sample.















Table 11






Fixed effects estimation results: Fishing revenue across weeks at lake Ahémé

Dependent variable:
Logarithm of fishing revenue for shrimp



Variables

(1)

(2)

(3)

(4)

(5)











Open first week

0.339*

0.507***

0.269*

0.311*

0.271


(0.187)

(0.185)

(0.162)

(0.180)

(0.183)

Open second week

0.116

0.457**

0.244

0.311*

0.281


(0.207)

(0.200)

(0.173)

(0.180)

(0.182)

Open third week

-
0.258

0.046

0.051

0.098

0.162


(0.260)

(0.263)

(0.201)

(0.212)

(0.202)

Open fourth week

-
0.533***

-
0.393**

-
0.165

-
0.168

-
0.098


(0.186)

(0.184)

(0.146)

(0.154)

(0.143)

Time trend


-
0.180***

-
0.171***

-
0.173***

-
0.184***



(0.038)

(0.031)

(0.034)

(0.034)

Log(Fishing days)



4.071***

2.988***

2.897***




(0.158)

(0.830)

(0.788)

Log(Persons fishing)




2.813***

2.812***





(0.601)

(0.576)

Log(Value fishing gear)





0.140**






(0.068)

Shrimp pots





1.071






(0.694)

Constant

7.016***

8.303***

1.045***

0.871

-
0.263


(0.104)

(0.310)

(0.390)

(1.573)

(1.706)







Observations

1,652

1,652

1,652

1,518

1,518

R
-
squared

0.009

0.062

0.369

0.143

0.152

Number of individuals

118

118

118

118

118

Notes:
Coefficients are reported with individually clustered standard errors in parentheses. ***, ** and *
denote significance at the 1, 5 and 10 % levels respectively. Lake Nokoué and lake Porto Novo are excluded
from the regression sample.