Anarchy, State and Somalia
Reflections on the viability of anarchy and the consequences of privatized justice
Boudewijn R.A. Bouckaert†
Would life be better, would wealth be higher, would society be more just without a state? These
questions about the desirability of anarchism have been one of the most fascinating topics in
political philosophy. Anarchism, either in its classical leftist variant or in its libertarian variant, is
the most radical answer to the evils which can accompany state authority such as war, massacres
and genocides, tyranny, oppression, class exploitation. The lecture of Rummels’ book ‘Death by
Government’ may well instigate a quest for a society without a state as is pointed out that
government was the most lethal factor during the twentieth century (Rummel,1996).
Professor Hans-Bernd Schaefer can be hardly qualified as an anarchist sympathizer. In his
writings and researches he emphasizes relentlessly that stable, transparent, efficient and non-
predatory public institutions are a prerequisite for free and wealthy societies. His life-long
commitment to development economics and his support for law-and-economics research that may
foster the efficiency of public institutions in less-development countries, witness this deep
Discussing anarchism is however not in opposition to Hans-Bernd Schaefer’s intellectual quest.
In their seminal article on adjudication as a private good Posner and Landes show that the
analysis of the possible outcomes of a totally privatized justice may well lead to robust a
contrario - arguments in favour of at least some state involvement in justice provision (Landes
and Posner, 1979). The analysis of anarchic scenarios may shed some light on the right- or
wrongness of the several justifications of state power. In short, the study of anarchy may learn us
why ‘archy’ might be better and, if so, which ‘archy’ is better, and these questions fit entirely the
endeavours of our celebrated colleague.
The three first sections of this article are theoretical. Based on insights on game theory and
institutional economics the possible consequences of a full privatization of justice and the move
into full anarchy are analyzed.
In the first section we analyze the relationship of the privatization of justice with other
privatizations in order to define the world in which the move to anarchy is made.
In the second section the Hobbesian argument against anarchism is refuted while the possibility
of ordered anarchy is made plausible.
In the third section we analyze in how far ordered anarchy is stable or vulnerable to tendencies
pushing it towards a relapse in a new kind a statism.
The fourth section is entirely devoted to the Somalian case. As Somalia provides us with the only
present real world case of anarchy, it is interesting to analyze it in the shadow of the theoretical
discussion, developed in the three former sections. Our conclusion is very mixed as it is difficult
to purify statist conflicts, ideological drives and premodern elements from the Somalian case.
Section 1: Privatization of Justice and Anarchy
In libertarian literature the types of goods and services which should be privatized are implicitly
sequentially ordered. The sequence is apparently inspired by the rate of political feasibility of, or
resistance against such privatisations. By ranking these types according to this criterion we arrive
at the following order:
1. Private goods and services (excludable and rival): reliance on free market has wide
support among economists and is, since the last decennials of the twentieth century,
also prevailing in politics.
2. Public goods or club goods (excludable but non-rival) : reliance on voluntary club-
provision has support from some economists but is sparsely accepted in politics
3. Collective goods (non-excludable and non-rival): reliance on voluntary provision
(based on kin-solidarity, moral commitment, tit-for-tat- strategies, repeated games) is
only envisaged by an outer libertarian fringe among the economists and is generally
abhorred on the political stage. This is especially the case for collective good
provision involving the possible use of violence such as national defence and justice
provision. The privatisation debate is haunted here by the nightmare of Hobbesian
This sequential ranking has heavily influenced the structure of the present debate on anarchy. As
the privatisation of justice and defence evokes the fiercest intellectual and political resistance,
these collective goods are generally considered to be the last to be privatized. Consequently the
present debate on anarchy is routinely structured as a choice between a state, reduced to justice
and defence-provision, the so-called minarchy or ultra-minimal state (Nozick, 1974)) and a
society, in which these goods too are provided by private enterprises, competing on free markets.
This sequential ranking within the privatization-debate is linked with the order of intellectual
viability among economists and the ensuing political strategy of libertarians. It is however not a
It is quite thinkable that states are involved in types of activities which qualify in the mentioned
sequential ranking as easy-to-privatise (intellectually as well as politically), but not in types of
activities which are in the same ranking qualified as hard-to-privatise. Several historical examples
can be mentioned in support of this.
As far as redistribution is concerned we can imagine states, involved in wealth-redistribution
between social groups but leaving the adjudication of conflicts and even the enforcement of the
sentences to non-state agencies (clans, customary courts, private justice providers). The
institutional setting during the early middle ages in Europe fits more or less this picture of a mere
redistributive state. Kings, local rulers (dukes, counts, prince-bishops, etc.), feudal landowners
and clerical institutions exacted rents from the peasants, artisans and merchants but left the
adjudication to private and customary courts and the punishment of crimes to the mechanism of
the blood feud (Bouckaert, 1997; Berman, 1983, 49-84)
The same applies for the regulatory activity of the state. In the libertarian sequential ranking this
type of privatisation (p.ex. leaving regulation to private labelling agencies, to ‘private city’-
authorities) would come before the privatization of hard-core state functions such as justice and
defence. It is, however, thinkable again that states use their monopoly of violence to regulate the
behaviour of people living under their control while leaving adjudication and defence entirely to
non-state institutions. The Taliban-state in Afghanistan fits more or less this picture. Adjudication
remained an affair of clan authorities deciding on customary law. The Taliban-authorities were
particularly keen on policing the behaviour of people according to the strictest canons of Shari'a
-law (beards for males, no schooling for females, etc.).
It is also thinkable that state activity concerns quasi-exclusively warfare (defence and aggression)
while not involved in adjudication. Several empires fit this combination. The Mongol empire for
instance was a mere warfare state, leaving all other functions to local authorities. Less martial
empires on the other hand, such as the Roman, the Ottoman and the British ones, aimed at a
gradual integration of the local legal orders and the adjudication institutions into an all
encompassing legal system (ius gentium, Ottoman law and common law) (Deepak Lal, 2004).
The absence of a sequential necessity of privatisations does not mean that the order by which
privatisations are introduced is without any impact. The fact that one domain of social action
remains under state monopoly may well affect the outcomes within other domains of social
action, which are privatized indeed.
The outcomes within the realm of private justice in a world of state monopolized defence (‘world
A’) will be different from the outcomes in this realm in a world of fully privatized and competing
armies (‘world B’).
In world B the weaker parties or weaker private justice agencies will very likely confederate
against stronger ones. The stronger one may also look for confederates, which may lead to new
In world A a weaker party will very likely run to the monopolized state army in order to support
her against the stronger one. This may trigger off a take-over process by the army-organisation
ending up in a state monopolized justice system. The world A-scenario occurred more or less in
late medieval Europe. Monarchic warfare states, aiming at subduing local rulers, gradually took
over local justice provision in order to establish full sovereignty over the subjects. This was often
done at a request for help by weaker parties at the local level (Tilly and Blockmans, 1994).
Because the sequential order of privatisation is not neutral with regard to the outcomes within the
privatised realms, we will analyse the possible outcomes within the realm of private justice
within an anarchic world, completely devoid from state monopolies. This discussion will shed
light about the viability and desirability of two possible worlds: 1) full anarchism including
privatized justice 2) a minarchy, i.e. a state solely devoted to the adjudication of private conflicts
and the implementation of sentences. This huge simplification of the context should be kept in
mind when the analysis is carried to the Somalian real world example of anarchy.
Section 2: The Hobbesian argument against anarchism
The Hobbesian picture of a world without effective government remains one of the most popular
arguments in favour of state power. Remove the state and people will start endless fighting
against each other. This will jeopardize social cooperation and labour division. The economy will
sink into full autarchy. So we will end up in a situation where life is ‘solitary, poor, nasty, brutish
and short’ (Hobbes, Leviathan, 1651)
Although very popular and ‘Pavlovly’ repeated in the courses of public law and political theory,
this Hobbesian argument does not resist game theoretical analysis.
This analysis refutes the Hobbesian argument in three steps: 1.A one shot-confrontation game; 2.
A repeated confrontation game; 3. An arbitration game;
In a one shot-confrontation-game the private justice agencies Able and Tough have two
strategies, confrontation or surrender and face the following outcomes: mutual conflict,
domination, submission, mutual submission. In the following matrix these outcomes are tagged in
a cardinal way:
- total stake of the game : 12
- domination means taking 10 of the stake minus 1, i.e. costs of the thread : 9
- submission leads to giving up the largest part of stake, 10 of 12 : 2
- mutual submission means peace and equal split of the stake : 6, 6
Confrontation 0,0 9,2
Surrender 2,9 6,6
In this game S,C and C,S are the Nash - equilibriums. Neither efficiency (S,S) is reached nor
justice. Whether S,C or C,S will be reached depends on arbitrary circumstances such as who is
the first mover, who has the more credible threat. Nevertheless the game does not end up in
(C,C), the Hobbesian state of nature. War is too costly. Parties prefer even submission above war.
(Tyler Cowen and Sutter, 2005)
In a context of repeated games the submitted party can punish the other party by not playing
surrender in the next game by which the confronting party in the initial game faces the outcome
of submission or mutual confrontation. So it is likely that the parties will end up in (S, S), the
efficient outcome. (de Jasay, 1997, 192-212; Rutten, 1999)
Under anarchy it is likely that the private justice agencies will order their relationship even
further by resorting to arbitration. Instead of solving each interagency conflict by a confrontation
game, the parties can delegate the resolution of such conflicts to an arbitration agency (Cowen,
1992; Cowen and Sutter, 2005). Such an arbitration agency is only acceptable by a private justice
agency when it is free from systematic bias towards its customers. The best way to avoid such a
bias is to adopt open-ended criteria to solve conflicts between customers from different private
justice agencies. Traditional examples of open ended-criteria are: the Lockean first come, first
served rule for property conflicts, the negligence rule for conflicts on damages, and the good faith
rule concerning the execution of contracts. It is virtually impossible for a private justice agency to
predict whether its customers will be rather squatters than Lockean first comers, rather negligent
agents or due caretakers, rather mala fide or bona fide parties. As it is likely that the private
justice agencies will recruit a very mixed public with regard to these criteria, they will accept
arbitration under these conditions. It saves them costly confrontation situations, while systematic
bias against their clients will be avoided.
A first conclusion of this game theoretical analysis leads to the refutation of the Hobbesian
argument. Full private justice in an anarchic world will not lead to a war of all against all and to a
relapse into autarchy. We can expect rather peaceful interaction between private justice agencies,
the development of arbitration agencies and the development of open-ended rules to solve
conflicts, quite similar to the tradition of the common and the civil law.
Section 3: Is ordered anarchy stable?
Ordered anarchy may be possible and even likely, this does not mean however that is it also
stable. The Walhalla of ordered anarchy, emerging from the breaking up of the last state
functions, may well be temporary and submitted to dynamics, leading irresistibly to a new
‘archy’. Again game theoretical analysis, especially the one concerned with coordination games,
network-industries and path dependence yield us interesting insights on this matter.
Analytical conclusions on the stability of ordered anarchy may lead to important conclusions on
the normative question whether a move from minarchy to anarchy is recommendable. If ordered
anarchy is not stable and relapses necessarily into a new type of ‘archy’, we have to consider the
following three possibilities:
1) the post-anarchy-archy is worse than the pre-anarchy-archy : in this case society
incurred a double loss
a. the costs of the quality decline of the archy
b. the costs of the double transition archy-anarchy-archy.
In this case a move to anarchy is not recommendable.
2) the post-anarchy-archy has the same quality as the pre-anarchy-archy: in this case
society incurs only the costs of the double transition
In this case a move to anarchy is not recommendable.
3) the post-anarchy-archy is better than the pre-anarchy-archy: when the improvement of
the archy is larger than the costs of the double transition, a case can be made for
breaking up the minarchy. The transition from archy to anarchy and back to archy has
in this case a purifying effect. The double transition costs have to be considered as a
macro-social investment in the building of better institutions.
The discussion about the viability of ordered anarchy took an interesting turn since the Tyler
Cowen advanced his so-called impossibility-theorem. (Cowen and Sutter, 1999) This theorem
signals a paradox in libertarian thinking related to the problem of collective good-provision.
Concerning the provision of ‘good’ collective goods libertarians argue that state involvement is
unnecessary, even harmful, because these goods will be more efficiently produced through
voluntary action. Libertarian literature refers here to the potential of moral capital, repeat playing
and procedures such as money-back-guarantees. Libertarians believe that the cooperative
potential within a society is strong enough to overcome the prisoners’ dilemma hurdle of
collective good provision.
Concerning the problem of collusion leading towards price-fixing cartels, libertarians however
point to their inherent instability due to the free riders’ attitudes of the colluding members.
Towards the colluders the stability of the cartel is a collective good. From a social point of view
such a price cartel is of cause to be considered as a ‘bad collective good’. Mainstream economic
thought agrees unisono that this bad requires a vigorous repression through anti-trust-regulations.
The libertarians however disagree. Cartels, they say, are unstable because members of the cartel
will be tempted to sell under the fixed price in order to increase their market share. Often they
will use all kind of cheating devices such as hidden price reductions or quality increase. Because
cartels are inherently unstable anti-trust-regulations are not justified. (Armentano, 1990)
According to Tyler Cowen, the libertarians want to have the cake and eat it.
Either you perceive strong cooperative potential within society allowing you to argue for a
withdrawal of the state from good collective good-provision but then you have to accept also
state involvement to prevent bad collective good-provision. For a strong cooperative potential
will also make collusion very stable.
Or you perceive only a weak cooperative potential in society and a predominance of free riders’
attitudes. In this case you can bet on the collapse of price fixing cartels and argue for the
redundancy of anti-trust-regulation but then you can expect problematic underproduction of good
collective goods, which makes state involvement in this realm a viable option.
Applied to the problem of privatization of justice the impossibility theorem leads to the following
dilemma. Either you are sceptic about cooperative potential in societies, which makes Hobbesian
anarchy likely as a result of privatisation of justice. Or you accept a strong cooperative potential
within societies, which makes it unlikely that privatisation of justice will produce Hobbesian
anarchy; but in this case the strong cooperative potential will likely produce collusion between
the private justice agencies, bringing us back to the state.
Tyler Cowen agrees that Hobbesian anarchy will be unlikely but points to crucial destabilizing
features of ordered anarchy. Private justice provision combines the characteristics of a network
industry with huge economics of scale (Cowen, 1994).
It is clear that the more clients a private justice agency has, the more attractive a subscription to a
private justice agency will be. The more clients, the less the likelihood of difficult interagency
conflicts. As a consequence the number of private justice agencies will tend to be rather small.
Moreover, the private justice agencies will set up a network of arbitration in order to avoid one
shot interagency conflicts. Stand alones, staying out of this all encompassing network, will face
grave difficulties. When their clients are involved in conflicts with clients of a network firm,
either a costly negotiation or a violent clash has to be expected. (Cowen and Sutter, 2005)
In addition, economics of scale may be considerable as private justice agencies may have to
invest big weaponry, in order to threaten off possible pirate firms in violent clashes. This big
weaponry can not be rented from another specialized firm as this firm would by this acquire a
dominant threat position.
As a consequence we may expect that privatisation of justice will bring about a strong network of
a small number of strong private justice agencies, which will tend to collude and limit
competition by fixing the prices. The network will gradually acquire the basic characteristics of a
state in the Weberian sense, i.e. an organisation exerting the monopoly of violence in a given
territory. In a more specific sense the state, coming out of a privatisation process of justice will
show the characteristics of a ‘shareholders’ state’. The major shareholders of the firms within the
network will indeed rule the country. (Cowen, 1992)
Caplan and Stringham however challenge Tyler Cowen‘s profound doubts on the stability of
ordered anarchy by stressing the game theoretical difference between networking, which is
essentially a coordination game, and colluding, which is essentially a prisoners’ dilemma.
(Caplan and Stringham, 2003) Networking between private justice agencies will be particularly
easy as there is no necessity for these agencies to choose between a standard of one or another
firm. For interagency conflicts they can opt for a mixed standard. In this case there are no
obvious winners or losers. Suppose private justice agency A adopted negligence liability as its
standard to solve intra-agency conflicts while private justice agency opted for strict liability. The
interagency conflicts, eventually brought before an arbitration agency, could be judged according
a mixed standard such as strict liability for unilateral accidents and negligence for bilateral ones.
As the investment in a certain legal standard is not dramatically high and can be easily corrected,
the possible lock in into a suboptimal standard is not a big issue here.
The stability of collusion, Stringham and Caplan argue, is a more difficult horse to ride as the
cartel will be under constant pressure from free riders, abating their prices in surreptitious ways.
Stringham and Caplan rely on some interesting historical examples such as the present credit card
industry and the clearing houses in the nineteenth banking sector. Both are examples of
successful and very encompassing networks maintaining however high competitive levels.
In response to these arguments Tyler Cowen points to the specificity of a market of privatized
justice.(Cowen and Sutter, 2005) Competition by outsiders who are not-integrated into the
arbitration network, may involve violent clashes, an eventuality clients will be eager to avoid.
Stand alones in the private market for justice face much higher costs than stand alones in, for
instance, the software market where programs such as Mac Os X Leopard are able to survive in a
world dominated by the Windows Network. For their users the inherent quality of the programme
is worth the costs of non-compatibility. Customers of justice agencies face a much harder choice:
sticking with their stand alone and risking violent clashes with the dominant network or swapping
to a network firm and enjoying the benefits of peaceful settling of interagency conflicts through
arbitration. As the dominant firms know that the network as such is relatively free from
competitive pressure and as it is very likely that the firms within the dominant are not very
numerous, a relatively stable cartel seems to be likely. Cheating will of course occur, but
probably not endanger the stability of the cartel.
Section 4: A glance at anarchy in action. The case of Somalia.
Until the nineties of the former century the theory of anarchy had no real counterparts in the
empirical world. Empirical cases were often sought in a very remote past, in the historical epochs
before the emergence of the nation state. Primitive and tribal societies were, for instance, studied
as examples of operational anarchies. (Benson, 1990) The empirical argument drawn from the
study of primitive society is weak however. The economic development of these societies barely
passed the phase of autarchy. Moreover, it is doubtful whether it makes sense to apply our
modern notion of individual liberty to social contexts, characterized by strong clan allegiances
and closed, not to say, isolated societies. The viability of tribal societies without state
organisation does not yield convincing arguments whether a developed economy, based on
sophisticated networks of cooperation, is viable under an anarchic institutional setting.
Somalia offers us the first real world example of anarchy since the world map has been fully
covered by the territories of nation states. The full coverage has been accomplished at the
Conference of Berlin in 1884-85, when the last non-state-governed spots on the world map have
been carved up by the most powerful nation states of that epoch, i.e. the European colonial super-
powers. Since then, the whole world was brought under ‘state-archy’. The evolution in Somalia
puts an anarchic spot on the statist-archic blanket of the world.
The Somali Democratic Republic gained, together with so many African states, independence on
1 July 1960. Its territory was a merger of British Somaliland and the Italian Somalian colony. By
a coup d’état in 1969 the country fell under the dictatorship of Siad Barre. Until the Ogaden war
with Ethiopia, Somalia was a staunch ally of the Soviet Union. Domestic policies were tainted by
Soviet recipes such as government control on land, a state planned economy, politics
monopolized by one party. The war with Ethiopia, which was backed by the Soviet Union, tilted
Somalia within the western camp. This however did not push the country towards more
liberalism and democracy.
In 1991 the regime of Siad Barre collapsed, but no political faction was able to gain control on
the whole of the country. The country fell apart in three parts. The North-Western part, in fact the
former British part, became an independent state called Somaliland, although not recognized by
the international community. The Northern part in the horn of Africa became a separate political
entity, called Puntland, not aiming however at the status of an independent state. In the southern
part, Somalia, no governmental organisation was able to establish itself in a stable way. In fact,
Somalia could be called, between 1991 and 2007, a genuine anarchy. The transitional federal
government, considering itself as the ruler of the whole of former Somalia and strongly backed
by Ethiopia and the US, was never able to gain any effective control and collect taxes. In 2006
the so called Islamic Courts wiped the warlords out of Mogadishu. From there they took control
over the largest part of Somalia and threatened even Baidoa, the last stronghold of the transitional
federal government at the Ethiopian border. Between 21 December 2006 and 9 January 2007, a
regular war was fought between the Islamic Courts and the Ethiopian army, backed in this by the
US. This led to the defeat of the Islamic courts and the occupation of Somalia by the Ethiopian
army. The Islamic Courts continue the fight as a guerrilla war. Foreign intervention is ubiquitous
in this evolution. Eritrea, Ethiopia’s archenemy, backs the Islamic courts, while the transitional
federal government is a phantom organisation, set up and heavily supported by Ethiopia. Guided
by the belief in strong Al-Qaeda presence in Somalia, the US has supported the secular warlords,
ruling Mogadishu before 2006.
In the following part of this section we will screen this contemporary real-world case of anarchy
though the prism of the theoretical analysis of sections 2 and 3. We will check, as far as possible,
whether 1) Hobbesian anarchy occurred or not, 2) the economy lapsed into autarchy or not, 3) the
anarchy was stable or tended to the re-establishment of state control.
Press reports and television programmes leave the impression that Somalia is in the grip of
constant and ubiquitous fighting between warlords and Islamic militias. Somalia is seemingly
Hobbesland. This impression is however false. Although Somalia is far from a heaven of peace
and is not a look alike Switzerland, the situation between 1991 and 2006 does not fit the
Hobbesian picture of the war of all against all.
To be fair, one should discount from the total level of violence the one which is due to foreign
intervention. This violence is not at the account of internal anarchy but has to be blamed to an
exogenous effort to end anarchy and to establish state control again. Such interventions were
often vigorous and led to a lot of bloodshed. Among others we mention the famous US-operation
Restore Hope in 1992, which triggered of a long series of anti-American ambushes, killing a lot
of Somalis and several American soldiers. In 1993 the warlord Aidid launched a raid to the
capital of Mogadishu to wipe out the Americans. Finally the Americans withdrew. In 1995 the
UN launched its operation United Shield. Again a failure, but with a lot of bloodshed. In 2006 the
Ethiopian army invaded the country transforming the loosely connected Islamic courts into a
more organised fighting force. Again the ensuing war took many lives, from soldiers as from
civilians. The pictures of these fights are prominent in press reports and leave the impression this
all is due to the internal situation of anarchy.
For most genuine internal violence the so called warlords of Mogadishu and surroundings are to
blame. The warlords, mostly related to a clan, controlled parts of the city, exacted taxes of the by-
passers for the use of roads, for ‘protection’, for the use of other public utilities such as water.
Internal violence was fierce shortly after the collapse of Siad Barre in 1991, but declined steeply
during the following years. The number of people killed by violence is high in comparison with
Western nations but not in comparison to other African countries, ruled by peace-bringing
governments. The number of deaths, killed by violence amounts up to 4 % of total deaths. This is
not higher than the deaths due to childbirth. In Mexico the number of violent deaths amounts up
to 3, 6 % of total deaths, a figure not much different from Somalia. (Leeson, 2006, 10)
The anarchic rule by the warlords and later by the Islamic courts did not lead to a predatory race
among these independent taxing authorities. The level of exaction, to be paid to warlord-militias
or Islamic courts is estimated at 5 %, while the fees for protection at another 5 %. This figure is a
lot lower than the taxation under government, domestic as foreign as well. (Leeson, 2006, 24)
Security was provided by the secular militias of the so-called warlords or by the militias
connected to Islamic courts in exchange for fees. Such fee –charging security firms are also
common in Western nations. In fact the difference with state government in this respect seems to
be rather gradual and not discretionary.
The judicial authorities continued to solve conflicts mainly on the base of customary law. The
Xeer-customs rather regulated marriage-relationships, the use of collective resources and the
social contracts between clans. The Diya-law dealt with serious criminal offences. Adjudication
was done by shari’a courts, almost always exclusively related to a clan. The courts were financed,
partially by donations from businessmen and by fees of litigants. (Leeson, 2006, 26; Van Notten,
2005; Nenova and Harford, 2004, 3)
Considering the continuous operation of these institutions and the limited amount of violence,
one can hardly depict the Somalian anarchic era as a Hobbesian war-situation.
Neither does the picture of autarchy apply. We assume state institutions such as a court system
and police to be vital for sustaining the broad network of cooperation, a viable market economy
requires. Absence of such institutions will make cooperation unstable, markets will fade out and
producers will relapse into autarchy. Also this dark picture does not fit the Somalian anarchic era.
In general, one can say that the economic situation improved considerably since the downfall of
Leeson compared key indicators of human development in the last years of the Barre-regime
(1985-1990) and during five years of anarchy (2000-2005). (Leeson, 2006, 12)
Table 1. Key development Indicators before and after Statelessness
1985-1990 2000-2005 Welfare change
GDP (PPP constant $) 836 600 ?
Life expectancy (years) 46.0 48.47 Improved
One year olds fully immunized against measles (%) 30 40 Improved
One year olds fully immunized against TB (%) 31 50 Improved
Physicians (per 100,000) 3.4 4 Improved
Infants with low birth weight (%) 16 0.3 Improved
Infant mortality rate (per 1,000) 152 114.89 Improved
Maternal mortality rate per (100,000) 1,600 1,100 Improved
Pop. With access to water (%) 29 29 Same
Pop. With access to sanitation (%) 18 26 Improved
Pop. With access to at least one health facility (%) 28 54.8 Improved
Extreme Poverty (%< $1 per day) 60 43.2 Improved
Radios (per 1,000) 4.0 98.5 Improved
Telephones (per 1,000) 1.92 14.9 Improved
TVs (per 1,000) 1.2 3.7 Improved
Fatality due to measles 8,000 5,598 Improved
Adult Literacy rate (%) 24 19.2 Worse
Combined school enrolment (%) 12.9 7.5 Worse
As Leeson states: ‘The data depict a country with severe problems, but one which is clearly doing
better under statelessness than it was under government. Of the 18 development indicators, 14
show unambiguous improvement under anarchy.’(Leeson, 2006, 13) The decline for the first
indicator (GDP), a key indicator though, is explained by Leeson by the inflated figures, firm
managers of the planned economy under Barre were providing, by the important output of
military hardware and by the huge influx of foreign aid during the Barre period.
Leeson also compares anarchic Somalia with the situation in neighbouring states. (Leeson, 2006,
Table 2. Somalia vs. its neighbours
% Improvement or Decline in Development Indicators between 1990 and 2005
Djibouti Ethiopia Kenya Somalia
GDP per capita (PPP) ---- + 15.5 -4.1 +?
Life expectancy (years) -15.4 + 9 -15.6 + 5.4
Adult Literacy (%) ---- ---- +3.7 -20
Infant Mortality rate (per 1,000) +16 +28.5 +7.4 +24.4
Pop. With access to improved water (%) +1.4 -4.3 +35.6 0
Pop. with access to improved sanitation (%) +3.8 +333.3 +7.5 +44.4
Telephone mainlines (per 1,000) +40 ---- +28.6 +1,150
Leeson is right by saying that this comparison rejects the hypothesis that Somalia would have
improved equally whether it remained under (its) government or not. Except for adult literacy
Somalia performs better than its ‘statist’ neighbours.
One of the main sources of Somalia’s economic progress is the explosive growth of its large-
scale livestock trade. During the years 1991-1999 the Somali-Kenya cross-border trade more than
doubled. After the severe drought of 2000 it declined, but soared again after the drought. The
cross-border cattle trade is facilitated by brokers (‘dilaal’) who certify for buyers that the sold
cattle is not stolen. If this certifying proves to be wrong the brokers are liable. Dilaal-fees did
not rise during the period of anarchy, which suggest that crime, especially theft of cattle, did not
rise during the anarchic period. This of course, was a crucial factor in the growth of this wealth-
component. (Leeson, 2006, 20)
Also with regard to monetary institutions, anarchy did not lead to collapse. As money is the
medium of exchange par excellence, an over-all accepted currency remains vital for the
continuation of the market process under anarchy. With the state also the central bank of Somalia
disappeared. Nevertheless exchange remained to be based on the Somalian shilling (Sosh), the
former currency. Old notes continued to circulate. On some occasions political actors printed
notes in order to fund their activities. In 1996 Aideed, a Mogadishu warlord imported new
shillings. In 1999 the Transitional National Government (the predecessor of the Federal
transitional Government) brought 30 billion shilling into circulation. This of course led to
considerable depreciation against the dollar. Nevertheless, depreciation rates of the shilling were
less important than under the Siad Barre regime. IN 1986 the rate was 110 Sosh against one US
dollar. At the end of the Barre regime the rate was 5700 Sosh. During anarchy the shilling eroded
further but much less: from 5700 in 1991 to 10 000 in 2000. One can hardly call this an optimal
monetary arrangement. The example shows however that a situation without an official monetary
constitution may be better than one with such a constitution. (Leeson, 2006, 21)
Finally, anarchy seemed to be compatible with the rise of a flourishing financial sector.
Numerous remittance firms, called ‘hawilaad’, funnelled enormous amounts of money from the
Somalian diaspora to the country. Access to the financial sector is much greater than under the
time of the Barre regime. (Leeson, 2006, 23; Nenova and Harford, 2004, 3)
As with regard to violence, security and law and order, also the economic situation does not
reflect the Hobbesian prediction of a relapse into autarchy. Relieved from numerous
governmental predatory obstacles, many business flourish better than they did under the Barre
regime. All business, facing relatively low levels of transaction costs, such as for instance mobile
telephones, were boosting under anarchy. Business and public services, facing high transaction
costs, such as water supply, were however trailing and suffered from the fragmentation of public
authority. Gradually also services, for the provision of which government support is generally
thought to be necessary, popped up in the anarchic hotbed of Somalia. This is the case for
education, a sector gaining momentum by the steady increase of formal schools, Koranic schools
and universities (Leeson, 2006, 26).
A last question concerns the stability of the more or less ordered anarchy in Somalia. Was it
submitted to an irresistible tendency towards the formation of a new statist system, as Tyler
Cowen would predict, or was it endogenously stable?
A first assessment of the evolution of institutions in anarchic Somalia seems to confirm the
prediction of Tyler Cowen.
In June 2006 the Union of Islamic Courts defeated the US-backed ‘warlords’, united in the so
called Alliance for Restoration of Peace and Counter-terrorism. Islamic courts are not merely
courts, but religiously inspired organisations, providing not only adjudication, but also police
functions such as patrolling in the city quarters, apprehension of suspects, and execution of court
sentences. In fact they are all encompassing justice and security firms. The Islamic courts
originated in Mogadishu but spread out gradually over the largest part of Somalia. Before June
2006 the courts were only loosely connected and ideologically quite diverse. Some were Salafite,
some Qutubi, some liberal, and some eclectic (Wikipedia, Islamic Courts Union). The Islamic
courts gained quick and massive support of the population as they restored law and order,
reopened the Mogadishu port and restored the property to the numerous refugees.
After June 2006 central institutions were set up to coordinate and to regulate the action of the
Islamic Courts. On one hand an Executive Committee was set up under the presidency of the
liberal sheikh Sharif, on the other hand an Advisory Committee (the ‘shura’) under the
presidency of hardliner Aweys. These bodies started to act as real legislative bodies by ordering a
ban on ‘quat’, a popular drug and by imposing the requirement that women, walking outside
should be accompanied by a male family member (Marchal, 2007, 6).
To what extent the setting up of these institutions heralded the incubation of a new state
organisation is however doubtful. Neither the Executive Committee nor the ‘shura’ were able to
build up sufficient authority towards the different Islamic Courts on the field and to exert control
at the grassroots of the movement. It seems that, at the contrary, many decisions were taken in
actions on the field, and accepted as a ‘fait accompli’ by the central institutions (Marchal, 2007,
6). This ‘de facto’ powerlessness of the central institutions rather feeds the libertarian contention
about the possibility of a stable and ordered anarchy.
As a consequence, the train of events does not allow us to decide which of the two outcomes
would have prevailed if Somalian anarchy would have had its free course. Moreover two
additional elements, ideology and clan-allegiance, have to be taken into account in order to
determine in how far the Somalian experiment comes to the support of the possibility or
impossibility of a stable and ordered anarchy.
First, the weight of the ideological factor within the Somalian evolution should be accounted for.
In the discussion about the stability of ordered anarchy the ideological factor is completely
exogenised. The actors in Tyler Cowen scenario are driven neither by anarchist beliefs nor by any
statist doctrine. They act in their proper interest and though an invisible hand-type evolution they
end up in a kind of statism, the participants did not envisage in fact. This was not the case in
Somalia. The leaders of the courts shared the idea of a strong and united Somalia (Marchal, 2007,
11). The setting up of the mentioned central institutions was less the outcome of colluding efforts
to establish a price cartel. It was as much driven by the idea to rebuild state authority. In addition
many militants of the Islamic courts, especially the Youth Leagues) were driven by religious
fervour and by their dreams to live once a ‘Talibanized’ Somalia. (Marchal, 2007, 7) Although
this element has been grossly overestimated by US foreign policy, it was certainly a factor to
push for strong central authority, able to implement stern religious precepts. When we discount
the ideological factor from the Somalian evolution, the anarchist cause is somewhat strengthened,
for without this ideological factor, centralisation would have been still weaker.
Clan-allegiance, on the other hand has remained strong throughout the whole Somalian saga.
As in many African societies the clan-organisation provides mutual help to its members protects
its members, while the members owe respect to the clan-elderly and are obliged to contribute to
clan-solidarity. The Islamic courts were for a large part based on these clan-allegiances. Only a
few, the religiously most radical, had a ‘trans-clan’-dimension. These clan-structures contributed
a lot to the failure of the centralising institutions of the Islamic Courts. Decisions of these
institutions were mostly explained, whether it was true or not, as clan-biased decisions. The
parties that did not prevail before these institutions always blamed it to their wrong clan-
allegiance and never to their eventual own misconduct (Marchal, 2007, 7). These persistent clan-
allegiances undermine considerably the relevance of the Somalian case for testing the modern
theories on anarchism. Although its members run around with mobile telephones, receive money
from American bank accounts, take the plane to Mekka and drive around in technological
advanced vehicles, social life in Somalia is still largely structured along premodern lines. It is, as
anthropologists like to call it, a prismatic society: a strange mixture of hypermodern and archaic
lifestyles and social structures. In this sense Somalia barely serves as a testing ground for modern
theories on anarchism. The cohesion of non-state protection agencies in Somalia relied largely on
strong blood-relationship and age-old allegiances to nomadic clans. They do not prove whether
such protection agencies would be as cohesive in a modern, individualistic and utilitarian society.
They even prove less whether such agencies would give in to the drive towards networking and
collusion, as supposed in the economic scepticism on ordered anarchy.
Thoughts with anarchist leanings were probably produced in the intellectual traditions from all
over the world. The writings of the Chinese philosopher Lao-Tsé for instance are tainted by mild
anarchist ideas. Only in the Western intellectual tradition however anarchism developed as a full-
fledged political philosophy. Starting with the writings of Etienne de la Boétie in 1576, anarchist
thought has inspired a lot of intellectual production and vigorous political action. The relative
success of anarchism in Western culture cannot be explained without referring to other deep
backgrounds in this culture. In the first place we should refer to the idea of a legal order centred
on individual rights. It makes sense to argue for the abolishment of state institutions when society
can fall back on a largely internalised legal order, based on respect for individual rights, freedom
of contract and the obligation to compensate for harm done. Anarchy will then resemble to a
‘kritarchy’, in which judges solve most social conflicts (Van Dun, 2005). It makes also sense to
argue for the abolishment of state institutions when supposing a universal feeling of solidarity
urging most members of society to cooperate and to share voluntarily with others. This is the
Kropotkine-Proudhon picture of anarchism. Both ideas, individual rights and universal solidarity,
are deeply intertwined with the classical Greco-Roman and Christian roots of Western
Against this background discussing anarchism remains a sound challenge to authority. It invites
to a constant effort to provide a firm intellectual justification for authority and it reflects the
attitude to accept only that level of authority for which reasonable arguments can be invoked. The
discussion about the possibility and stability of anarchy, as outlined in the first section of this
article, highlights the possible contributions by game theorists and economists to this debate.
The Somalian anarchy however was not the result of a deliberate reformist or revolutionary
movement inspired by anarchist ideas. The withering away of the state in Somalia was not
preceded by a thorough public debate on the legitimacy of state institutions. Anarchy resulted
from an accident, the collapse of a totalitarian regime and the inability of political factions to
reintroduce state authority. Somalia was anarchy without anarchism and anarchists. Somalia also
misses the background against which Western anarchist thought was able to develop, i.e. legal
universalism and/or universal solidarity. Its background is still largely determined by premodern
clan-thinking and religious fervour. In this sense it is difficult to evaluate the merits of anarchism
by the Somalian experience.
Nevertheless, the analysis of the Somalian anarchist era is not pointless. It shows that markets
and civil society will continue to function without state institutions on a level, generally higher
than expected. It also shows that not any state is better than anarchy. The Somalian anarchy is in
many respects better than the totalitarian ‘archy’ that preceded it and performs in many respects
also better than its neighbouring states. ‘I just want a government, any government will do’,
answered a Somali citizen to a BBC-reporter, covering the turmoil in 2004 (BBC News). The
analysis of the Somalian situation shows that this man is wrong. Some governments can do a lot
worse than no government. But this is not a proof of the opposite statement that no government is
always better than any government. More than providing us with definite answers on the viability
of anarchism, the Somalian case shows us that society does not depend on government but that a
limited government can improve society in a substantial, though not unlimited way.
Dominick T. Armentano, Antitrust and Monopoly. Anatomy of a Political Failure, Oakland, 1990
Bruce L. Benson, The Enterprise of Law. Justice without The State, San Francisco, 1990
Harold J. Berman, Law and Revolution. The Formation of the Western Legal Tradition,
Cambridge (Mass.), 1983
Boudewijn R.A. Bouckaert, ‘Between the market and the State: the World of Medieval cities’. In:
G. Radnitzky (Ed.). Values and the Social Order, Volume 3, Voluntary versus Coercive Orders,
Bryan Caplan and Edward Stringham,’Networks, Law and the Paradox of Cooperation’, 16 The
Review of Austrian Economics, 2003, 309-326
Tyler Cowen, ‘Law as a public good: The Economics of Anarchy’ 8, Economics and Philosophy,
Tyler Cowen ‘Rejoinder to David Friedman on the economics of Anarchy’, 10 Economics and
Philosophy, 1994, 329-332
Tyler Cowen and Daniel Sutter, ‘The Costs of Cooperation’, 12 Review of Austrian Economics,
Tyler Cowen and Daniel Sutter’ Conflict, Cooperation and Competition in Anarchy’, 18 Review
of Austrian Economics, 2005, 109-115
Anthony de Jasay, ‘Against Politics: On government, anarchy and order’, Lectures Max-Planck
Institute for Research into Economic Systems, Jena, 1995
Thomas Hobbes, Leviathan, 1651
Deepak Lal, In Praise of Empires-Globalization and Order, New York, 2004
William L. Landes and Richard A. Posner, ’Adjudication as a Private Good’, 8 Journal of Legal
Studies, 1979, 238-256
Peter T. Leeson, ‘Better off Stateless: Somalia Before and After Government Collapse’,
Department of Economics, West Virginia University, 2006
Roland Marchal, ’Somalia: A New Front Against Terrorism’, Working Paper 5 February 2007
Tatiana Nenova and Tim Harford, ‘Anarchy and Invention. How does Somalia’s Private Sector
Cope without Government?’, Public Policy for the Private Sector, The World Bank Group, Note
nr. 280, November 2004
Robert Nozick, Anarchy, State and Utopia, Princeton, 1974
R.J. Rummel, Death by Government, 1996,New Brunswick and London
Andrew Rutten, ‘Can Anarchy save us from Leviathan ?’, The Independent Review, III, Spring
Charles Tilly and Wim P. Blockmans, Cities and the Rise of States in Europe A.D. 1000-1800,
Frank Van Dun, ‘What is kritarchy’, Appendix to Michael Van Notten, The Law of the Somalis.
A Stable Foundation for Economic Development in the Horn of Africa, Trenton NJ, 2005
Michael Van Notten, The Law of the Somalis. A Stable Foundation for Economic Development
in the Horn of Africa, Trenton NJ, 2005
Wikipedia, ‘Islamic Courts Union’, Paper last modified 19 October 2007