candle. In its place came the principle of national sovereignty: what-
ever wrongs a government perpetrated on its own population, they
did not have sufficient consequence upon the well-being of other
countries to warrant intervention. At the time the concept was de-
veloped in the seventeenth century, there were reasonable grounds
for such a proposition: economies and societies were not highly inte-
grated. Whether or not it was true at the time, it certainly is not true
any longer. Nowadays a civil war generates externalities for neigh-
bors that are too large and too adverse to be dismissed.
I have tried to measure them through studies with Anke, Lisa
Chauvet, and Alberto Behar. The approach we used was standard,
although care has to be taken to distinguish those neighborhood ef-
fects that have nothing to do with war from war itself. For example,
a neighborhood might be affected in common by a drought, as in
Southern Africa during the mid-1990s. We find, unsurprisingly,
that the costs of a civil war to any particular neighboring country
are considerably less than the costs to the country itself. Typically,
a country might lose around 0.9 percentage points off its growth
rate if one of its neighbors is at war. However, the typical civil war
country has three or more neighbors, and, further, the economies
of the neighboring countries are usually larger than that of the civil
war country itself. This is because, as we have seen, being small and
being poor are both risk factors.
In our analysis we include only costs to immediate neighbors.
This omits demonstrated adverse spillover effects across a wider
area. Even with the restriction to immediate neighbors, the num-
bers imply that the costs to the neighbors as a group are likely to
be even larger than the costs to the country at war. So, reflecting
the standard economic solution to the problem of how externalities
should be internalized into the decision process, decisions that sub-
Better Dead Than Fed?
223
stantially affect the risk of a civil war should be internalized among
the neighborhood. In the Democratic Republic of the Congo sev-
eral of the neighbors got involved: indeed, three of them, Rwanda,
Uganda, and Angola, sent their troops into the country. The neigh-
borhood dimension of security could scarcely be more graphically
illustrated.
Recall that by far the most dangerous situations are post-con-
flict. Post-conflict relapses are likely and inflict high costs upon
neighbors. Should the neighbors of a post-conflict state have some
rights to a say in post-conflict policies? That was where my thinking
had reached a year ago: post-conflict countries should go through
a phase of sharing sovereignty with their neighbors until they had
progressed out of danger. And then I woke up to two insuperable
problems.
Problem number one: Not only do neighbors have a legitimate
interest in the governance of the post-conflict country, they are also
likely to have some interests that are rather less legitimate. Around
the world, neighbors often have problematic relationships: after all,
they are overwhelmingly the main source of external threat. Paki-
stan, which as I write is imploding following the death of Benazir
Bhutto, is not going to share its sovereignty with India; Eritrea is
not going to share its sovereignty with Ethiopia. So neighborhood-
shared sovereignty is not going to work. The African Union rec-
ognized this when it proposed that the African peacekeeping force
for Somalia should be composed of forces from any willing coun-
try other than a neighbor. However, Somalia also demonstrated the
limits of that approach: the only country with a sufficiently strong
interest to send a major force was neighboring Ethiopia.
Problem number two: The neighbors of a country that becomes
post-conflict are not a natural political grouping. As a result they
have no experience of cooperating as a group. Worse, their coop-
eration would clearly be time-limited: it may last for only a decade.
Worse still, there may be rather a lot of neighbors. Take the Dem-
224
WARS, GUNS, AND VOTES
ocratic Republic of the Congo, which is currently a post-conflict
country, and look at a map. What grouping of neighbors do we get:
Congo-Brazzaville, the Central African Republic, Sudan, Uganda,
Rwanda, Tanzania, Zambia, and Angola. Experimental games find
that cooperation gets harder as the number of players is increased,
and eight participants is a lot. They also find that the players grad-
ually learn how to cooperate, so that starting up as a group from
cold would impose a phase of mistakes just when the post-conflict
country is most vulnerable. Finally, one of the most basic results of
experimental games is that players evolve a tit-for-tat strategy that
enforces cooperation: players avoid unreasonable decisions because
they would eventually get their comeuppance. So temporary coop-
eration is much harder than permanent cooperation.
These problems persuaded me that sharing sovereignty with
the neighbors is out. What then is the alternative? The solution, I
think, is to place the legitimate interests of neighbors in trust with a more permanent grouping that does not itself have strong interests. While this might be a regional body such as the African Union,
the more obvious locus is the United Nations, and more specifically
the Peace-Building Commission, which was established in 2005
and is jointly under the Security Council and the General Assem-
bly. So, in the case of the Democratic Republic of the Congo, the
United Nations would hold a share of sovereignty on behalf of the
neighbors, tasked with minimizing the shared costs inflicted on the
neighbors. To be clear about this, the United Nations would not,
in its own right, hold a share of sovereignty. This is far short of the old model of United Nations trusteeship that some scholars have
suggested should be revived. The post-conflict government would
share sovereignty rather than be stripped of it, and the objectives
of the regional or international body with which sovereignty was
shared would be predefined to be the protection of the legitimate
interests of neighbors.
What would guide the decisions of the trustees? To an extent,
Better Dead Than Fed?
225
each decision has to be based on the totality of the circumstances that
make each decision unique. But it would help to have an explicit set
of guidelines. Guidelines are useful where different players have to
be coordinated. I think that there are three distinct players. Some
governments should provide or finance peacekeepers; some govern-
ments should provide aid; and the post-conflict government should
reform economic policy, cut its military spending, and, if it chooses
to hold elections, let them be free and fair. The embargoes on im-
ports of armaments that are appropriate in post-conflict situations
have routinely been broken by rogue companies in countries that
have to date been below the radar screen of international scrutiny,
but we now have the means to identify such breaches.
Each of these depends upon the others. Peacekeepers are less
likely to get killed where arms embargoes are effective. The realis-
tic exit strategy for peacekeepers is economic development. In turn,
economic development is enhanced by aid and policy reform. Elec-
tions as commonly practiced to date, which is to say far from free
and fair, have increased the risk of violence, not reduced it. Not only
are these decisions interdependent, but they need to be sustained for
around a decade, whereas to date all three players usually focus only
on the short term. Guidelines that set out the mutual responsibilities
of all players over the course of the decade could not be legally bind-
ing, but they could create a common expectation of behavior. They
are also very much in the spirit of modern international coopera-
tion: from the Monterrey Consensus to the United Nations Global
Compact with large corporations, the approach has been to spell
out mutual responsibilities. As the inclusion of the responsibility to
comply with arms embargoes demonstrates, the responsibilities ex-
tend broadly and are not polarized between the governments of the
rich world and those of the bottom billion.
In setting out guidelines for the behavior of each party, a post-
conflict compact would also, implicitly or explicitly, reveal the red
lines that should not be crossed. The clearer the red lines, the less
226
WARS, GUNS, AND VOTES
likely they would be breached, and so international involvement in
post-conflict situations would become less of a nightmare. If some
red lines had been in place, I think that the post-conflict nightmares
might have been avoided.
I h av e p u t f o r wa r d t h r e e proposals for international action.
Between them they address the abuse of democracy in the acquisi-
tion of power, the misuse of power once acquired, and the structural
insecurity that has beset the societies of the bottom billion. Might
they be adopted?
At present, discussion of international action ranges across
the extremes. Think of the different stances on Zimbabwe. At
one extreme there have been calls both from within Zimbabwe,
by the Archbishop of Bulawayo, and from a range of international
commentators, for international military intervention to depose
President Mugabe. Tony Blair vetoed Mugabe’s attendance at the
Commonwealth Conference, and Gordon Brown refused to at-
tend the Africa-Europe Summit because Mugabe was included.
At the other extreme has been the indignant solidarity of African
presidents, manifested in the election of Zimbabwe as chair of the
United Nations Human Rights Committee. The three proposals
in this book are a very long way from using military intervention
for regime change. I think that externally imposed regime change
tramples on the unhealed wound of colonialism and so is unreal-
istic. They are also far from noninterference. In an interconnected
world, untrammeled national sovereignty leads unswervingly to
hell. The proposals are a compromise between positions that are
currently deadlocked.
If they were adopted, would they make a difference?
The disaster unfolding before my eyes as I finish this book is
Kenya. As the book has built up I have tried to show how the rule to
bolster the conduct of elections would quite probably have changed
Better Dead Than Fed?
227
the course of Kenyan history. But the disaster that has overshadowed
Africa for the past decade has been Zimbabwe. Manifestly, Presi-
dent Mugabe has systematically dismantled both the democratic
polity and the economy of his country. So what might have averted
this disaster? The only power that might realistically have changed
the course of Zimbabwean history is its own military. The African
Union now has a rule refusing to accept coups as legitimate. While it
is entirely understandable that incumbent presidents would happily
agree to such a rule, it is misplaced. Zimbabwe needed a coup, but
not one that led, as in Cote d’Ivoire and Ethiopia, to further ruin.
Coups need to be harnessed, not eliminated: the core proposal of
this book.
C h a p t e r 1 0
O N C H A N G I N G R E A L I T Y
We have come full course. The societies of
the bottom billion are structurally insecure and
structurally unaccountable. Despite recent years be-
ing the most successful period of economic global growth on record,
the appalling consequences are apparent to all. Structural insecurity
hit the headlines in 2007 first due to Somalia and then to Sudan.
Structural lack of accountability in the conduct of elections hit the
headlines in 2007, first in Nigeria, then in Pakistan, shortly followed
by Kenya. The year 2008 started with a rebellion in Chad and a coup
attempt in East Timor, the president of which is currently recuper-
ating in Australia. I fear that there will be many more such events.
So what, in a nutshell, is the structural problem faced by the
countries of the bottom billion? It is that they are too large to be na-
tions yet too small to be states. Too large, because they lack the cohesion needed for collective action. Too small, because they lack the
scale needed to produce public goods efficiently. Societies can func-
tion well enough without some public goods because they can also
be supplied privately. Some of the health and education services that
are supplied as public goods in Europe are supplied as private goods
in America. But some other public goods cannot be substituted by
private activity. Security and accountability are such goods.
230
WARS, GUNS, AND VOTES
No society can rely successfully upon private security, although
periodically it has been tried. The private forces hired for defense
become predatory on the very people they are supposed to protect.
Britons tried it after the Romans departed, hiring a gang of Jute
thugs to defend them against the Picts. It took the thugs fifteen
years to work out the obvious: they then slaughtered the British po-
litical elite and took over. As for the private provision of account-
ability, most instances that appear to be private provision, such as
the way that the American health system is disciplined by the fear
of being sued, depend upon being backstopped by the rule of law.
In the absence of the rule of law the need to maintain a good repu-
tation within a small network of associates can enforce a degree of
accountability. Economists regularly parade the example of how
thirteenth-century Jewish traders conducted long-distance trade
despite the lack of law. But Avinash Dixit has recently shown that if
such networks are scaled up the whole system crashes. Security and
accountability are either provided by government or they are not
provided. Their absence produces socioeconomic conditions such as
the botto
m billion have lived through for forty years. During that
time they have become the poorest people on earth.
With sufficiently visionary political leadership, the states of
the bottom billion could build a shared identity within the society,
thereby transforming state into nation, and cooperate with the other
nations of their region. Combined, these approaches would enhance
the supply of the public goods, providing the security and the checks
and balances that their citizens need. From time to time people ca-
pable of such leadership gain political power, but not very often. It
is not by chance that the visionary leaders Julius Nyerere, Sukarno,
and Nelson Mandela were all founding presidents. Once political
power can readily be won by the self-serving, the self-serving will
step forward to try their luck and the honorable will step back. Bad
currency drives out good. In this book I have spared you the fancy
terminology of economics, but since you have reached the end you
On Changing Reality
231
can take delight in one technical term: in economic language the
quality of political leadership is endogenous. As a result, in these
societies visionary leadership is now rare.
There is thus a powerful case for security and accountability to
be regarded as basic social needs that, as a default option, should be
provided internationally. After the intervention in Iraq, many peo-
ple might reasonably feel that the unintended consequences of secu-
rity interventions are such that intervention in any form is too risky.
Yet international military intervention has had many successes. The
lesson is not that it is intrinsically risky, but that the circumstances
that warrant it should be limited and clearly delineated.
Since any such proposal will be met by a chorus of outrage from
the beneficiaries of presidential sovereignty, the five billion who live
in territories that have been more fortunate can readily justify their
natural proclivity to stand back and watch. It will also be the con-
clusion of those in thrall to a sense of victimhood: the rest of the
world has already done enough damage. My own attitude used to
be “just give it time.” After all, in the countries that are now devel-
oped, the transition from the effective but unaccountable state of the
Paul Collier - Wars, Guns, and Votes Page 28