Writing Well

William Strunk Jr: “The Elements of Style” and Bryan A. Garner “HBR Guide to Better Business Writing”

Two very insightful books on how to write well. Perhaps a dry subject, but also an important one. Rather than critically reviewing the two books, I here aim to provide a “cheat sheet” of Strunk’s and Garner’s most pertinent recommendations.

First published in 1918, Cornell University Professor William Strunk's book has become a classic. This English professor's book offers a lot, but, above all, his “elementary principles of composition” are timeless:

• The paragraph is the basic unit of composition. Each paragraph, in turn, should have one topic. With a new paragraph, the writer signals to the reader that he/ she developing a new topic.
• Each paragraph should have a beginning, a main part, and an end. A paragraph should begin with a topic sentence describing the paragraph's purpose. The writer may then proceed to develop the topic in several forms, he "may make the meaning of the topic sentence clearer by restating it  in other forms, by defining its terms, by denying the converse, by giving illustrations or specific sentences; he may establish it by proofs; or he may help develop it by showing its implications and consequences." Finally,  a paragraph's final sentence should either put an emphasis or state a consequence.
• Use the active voice because it is more forceful (well, unless the passive voice is necessary or more convenient). Strunk points to two cases in which the passive voice should definitely be avoided: two passive constructions should not depend on one another (“Gold was not allowed to be exported”) and what was originally a passive construction should never become the subject of a sentence ("A survey of the region was conducted in 2001").
• Put statements in positive form. Otherwise, the writer risks producing colorless and non-committing language. It is better to say “dishonest” instead of “not honest”; “trifling” instead of “not important”; “forgot” instead of “did not remember”; “ignored” instead of “did not pay attention”; “distrusted” instead of “did not have much confidence”.
• Omit needless words. "Used for fuel purposes" should be "used for fuel; "in a hasty manner” should be hastily", etc. Strunk also argues against the expression "the fact that". "In spite of the fact that" could for instance be better expressed as "though” or “although”, while "the fact that he had not succeeded" better as "his failure". 
• Avoid successions of loose sentences only connected by “and” or “but”. 
• Express coordinate ideas in similar form. Expressions do not constantly need to vary. Instead, it may be appropriate to use parallel constructions to express similar content and function. An example of parallel construction: "Formerly science was taught by the textbook method, now it is taught by the laboratory method.” Importantly, this means that correlative expressions such as “both ... and”; “either ... or”; “first, second, third" etc. should be followed by the same grammatical construction.
• Keep related words together. Subject and verb of a sentence should not be separated (e.g. “God, by virtue of his powers, judges”). Moreover, relative pronouns should follow immediately after their antecedents ("in his eyes was a look that" rather than "there was a look in his eyes that"), so to avoid confusion what those pronouns refer to.
• In summaries, authors should keep one tense. Using multiple tenses creates, as Strunk puts it, "the appearance of uncertainty and irresolution". Antecedent action is expressed in the perfect if the summary is kept in the present and in the past perfect if the summary is in the past. Also, expressions like "he said/ added/ expressed" are of no use, once the reader knows that he/ she is reading a summary.
• Place the emphatic words of a sentence at the end. This usually means placing the new element of a sentence- usually its predicate- at the end. Subjects can also be emphatic when placed at the beginning; however, to receive proper emphasis, subjects must take the place of the predicate ("Through the middle of the valley flowed a winding river"). 

Equally usefully, but perhaps a bit more modern and application-oriented in approach is Bryan Garner's "Better Business Writing". At the core of Garner's book stands his distinctions between four different stages of the writing process:

• The "madman" gathers materials and generates ideas.
• The "architect" takes what the "madman" has gathered and organizes it into an outline. Garner recommends for the “architect” to first generate a list of topics and then organize each topic into sets of three points.
• The "carpenter" puts together the text with its sentences and paragraphs based on the “architect's” plans. The “carpenter” should write as quickly as possible rather than waste time on achieving perfection.
• The "judge" conducts the quality control and will have to play first a revising, then an editing role. When revising, the judge should reconsider the text written in its entirety and evaluate whether it comes to the point and is fair, complete, truthful; whether there is an opener, a middle, and a closer; whether points are proven with specifics; whether headings are informative; and whether the closer is consistent, but fresh. The editor, in turn, enquires if there are better or more interesting ways of expressing an idea; if the overall expression could be described as relaxed, yet refined; and if sentences neatly glide into one another.

Similar to Strunk, Garner also lays out his own “principles of composition”. 

• Clarity is best achieved by keeping it simple- a sentence should not be any longer than 20 words-and persuasive- the text should offer evidence rather than "pre-cooked” conclusions.
• Brevity. Vital information should be summarized at the beginning of a document. Each section, in turn, should begin with a sentence summarizing the 5 w's (“who, what, when, where, why”).
• Delete superfluous wording. Every proposition possible, especially "of", should be replaced. Verbs should substitute for every "ion" noun whenever possible ("protected" instead of "provide protection"). Writers should also replace "to be" with stronger verbs and eliminate padding such as "in terms of" or "the purpose of".
• Avoid business speak. Just like when speaking, we should use language as direct as possible. Garner advocates using the Flesch Reading Ease scale. This scale relies on the length of sentences and syllables to derive what level of education somebody needs to understand a text. Developed as part of the “Plain English” movement, the scale measures a text’s accessibility from 0 to 100. Higher values signify that a text is more understandable. Texts that score above 60 can be called as written in plain English. Microsoft Office offers a built-in Flesch scale.
• Use chronology when delivering a factual account.
• Establish continuity through well-joined sentences. Garner offers a whole series of connecting words which allow to add ("further"), underscore ("above all", "chiefly"), concede  ("admittedly", "doubtless", "granted") or return to a point ("even so", "nontheless"); but also to set up conclusions ("on the whole") and contrasts ("conversely"). Also, subheadings may come in handy to establish continuity.
• Learn correct grammar.  Various recurring grammar and wording which tend to vex writing: subject-verb disagreements (“there is risk and danger considerations”); noun-pronoun disagreements (“a citizen may cast their vote”); double negatives (“I couldn’t help but); and non-standard English (“where is the meeting at”).  

Garner also discusses approaches how to avoid annoying one’s readers off.

• Avoiding to anesthetize one’s readers: Garner advocates the use of personal pronouns to pull the reader into a document and the use of contractions where possible (e.g. “remember:” instead of “We would like to remind you”). Further, language should be simple, but varied. Passive voice is to be avoided, and the writer should vary the length and structure of sentences. Acronyms tend to put readers off.
• Watching one’s tone: Hyper formality, sarcasm, and a non-collegial tone all have the potential to offend one’s readers.

Garner concludes his book with recommendations for various forms of business writing, to include memos, e-mails and letters. His suggestions for memos, in particular, stood out to me:

• Develop a short and clear title.
• Summarize key specifics up front.  A memo should start with a summary, which states the issue, the recommended solution, and the reasoning behind it in plain terms. Writing a memo also means constantly readjusting body text and summaries so they are in synch with each other.
• Focus on three audiences at the same time: executives, subject matter experts assessing the soundness of your document, and future readers, who, sometime down the road, may be required to understand your memo .

In conclusion, and just for the fun of it: this text scores 54.8 on the Flesch Reading Ease scale. This score means that the text is "fairly difficult to read" and at 10th to 12 grade school level. I used the passive in 10 percent of the sentences, intentionally so and just for the fun of it in at least two sentences. The average sentence length is 17 words.

The World is Flat

Richard Baldwin The Great Convergence- Information Technology and the New Globalization

Economic globalization since the late 18^th century explained in a highly insightful and immensely readable book. Breath-taking in its theoretical ambition and its empirical richness.

At the core of Baldwin’s book is the idea that there were two different globalizations which took place since the late 19^th century. He introduces those two globalizations in the first part of his book.

A first globalization in the 19th century was triggered by the massive lowering of transportation costs through the introduction of steam power. Railroads and steamships allowed to move goods at a much faster rate and reduced costs. This first globalization caused production and consumption, which had been co-located since the beginning of humankind, to geographically un-bundle. In a global perspective, the first globalization produced urbanization and concentrations of factories in Western Europe, but deindustrialization in traditionally productive areas such as China or India. A massive divergence thus set in between the Northern and the Southern hemisphere. Temporarily set back by two world wars and the world economic crisis inbetween, globalization took up speed again after 1945. The General Agreement on Tariffs and Trade, or GATT, locked in the principles of a rules-based world trade system. Another key technological factor lowering shipping costs and hence promoting trade has been containerization. Shipping in standard-sized steel containers has become increasingly far-spread since the 1960s. It avoids the delays historically faced by the loading and unloading at ports and focuses activity at the harbors on giant cranes moving standardized boxes. 

A second globalization begun in the late 1980s and allowed for an unbundling of previously tightly integrated production processes. Firms began to offshore whole production stages across borders. The revolution in the Information and Communication Technologies (ICT) and the growing importance of air cargo were key factors in this second unbundling. Three “laws” help us understand the ICT revolution: Moore's Law indicates that computing power grows exponentially, with computer chip performance doubling every 18 months. Gilder's Law that bandwidth grow three times more rapidly than computing power, doubling every 6 months. Finally, Metcalfe's Law postulates that the usefulness of a network rises exponentially with the number of its users. In summary, ICT spread at a revolutionary speed and facilitated the second globalization. Another key factor in the second unbundling was the spread of air cargo since the mid- 1980s. While air cargo remains significantly more expensive than other means of transport, it is both speedy and reliable and thus offers international production networks certainty they can receive intermediate goods they critically depend on.

The second unbundling resulted in a shift of manufacturing, to include jobs, from the traditional industrial powerhouses, the G7, to what Baldwin calls the Industrializing 6- China, Korea, India, Indonesia, Thailand and Poland. Those 6 countries account for an increasing share of the world's GDP. Trade between the I6 and the traditional manufacturing giants, the United States, Japan and Germany has grown significantly along the lines of international production chains. At the regulatory level, the emergence of international production chains was eased by increased protection of Foreign Direct Investment, particularly through Bilateral Investment Treaties.

The second part of the book aims to provide a more in depth understanding of the drivers behind the two “unbundlings”. Baldwin's model of globalization rests on three different types of costs wich affect production processes: the costs of moving goods through trade; the costs of moving ideas through ICT; and the costs of communicating face to face by moving people. 

• The first globalization entailed a massive reduction in transportation costs, allowing for the unbundling of production and consumption. This led to the rise of long-distance trade, with nations specializing along the lines of their comparative advantages. However, somewhat paradoxically, while production and consumption became unbundled and long distance trade rose in the 19th century, production processes tended to cluster. The reason behind this can be found in the theory of the New Economic Geography: Clusters offer the advantage of scale economies and particularly faster innovation by bringing together different people working on similar problems. The types of cluster arising in one nation, in turn, determined what the nation would be exporting. 
• The second globalization, in turn, was driven by a massive reduction in ICT costs. Production processes no longer had to be concentrated in one location, but could be distributed across borders. Still, with face to face coordination remaining high, offshoring remained regionally limited, usually to places which could be reached in a day of travelling. For example, German manufactures offshored to Poland, US manufactures to Mexico, and Japanese manufacturer to China and South Korea and later Vietnam.

In Chapter 5 of his book, Baldwin attempts to answer the question what difference the second globalization made. Production processes changed fundamentally during the second globalization and became organized across borders rather than in single factories. Trade diversified away from the previously dominant trade in goods and came to include trade in components, the movement of entire production facilities and the services needed to facilitate cross border production. Another fundamental change in international manufacturing since the 1980s is described by the so-called smile curve. Production chains can be split into three different areas: pre-fabrication, to include design or finance, fabrication or the actual manufacturing activities, and post fabrication, which includes sales and marketing. The smile curve indicates that since the 1990s, the relative value added of pre- and post-fabrication services has increased. The reason behind this is the decrease in production costs associated with this second stage of production. Another consequence of the second unbundling is that in industrialized countries, blue collar workers in the middle-income range were most negatively affected by offshoring. On the contrary, workers in service industries were not affected by off-shoring, while workers in knowledge-intense professions tended to profit. Overall, it is important to realize, though, that the new globalization has more individualized and less predictable effects than the old globalization: the effect of international competition is no longer felt sector by sector, but rather production stage by production stage and sometimes even by different jobs.

In Chapter 6, Baldwin introduces four different economic theories to help him account for the second globalization (or unbundling)'s impact.

Comparative advantage: David Ricardo's theory of the comparative advantage indicates that as long there is a difference in relative prices for goods between nations, market participants stand to gain from trade. In practice, opening to trade means a rise in the prices of goods a country is good at making and hence exporting; and a fall in prices of the goods a country was less good at making and hence importing. Trade openness may result in a more efficient industrial structure, with firms attempting to merge in search of economies of scale. Ricardo’s theory also has something to say about the winners and losers of globalization and the resulting societal conflicts. On one hand, increased trade openness may pit consumers and producers against each other: imports of a certain good, say cotton, benefit consumers, as they will lower prices, but will also hurt the domestic producers of cotton. On the other hand, globalization may also put whole sectors which stand to gain from enhanced trade openness against those which stand to lose.

New Economic Geography: Baldwin introduces the logic of the “new Economic Geography” to explain firm’s location choices. Theories of this school distinguish between forces dispersing economic activity and those driving its agglomeration. Wage gaps are one major dispersion force - skill intensive industries such as the IT industry need high-education labour usually situated in high-wage nations, while the opposite holds for labor-intensive industries, such as textile.  Local competition also has a dispersive effect: firms try to settle in locations where they face less competition for their products. On the agglomeration side of things, the New Economic Geography school points to the effects of circular causality (or what economist Albert Hirschman had called backward linkages). Firms have incentives to move to places where there are already other firms, both to be near to potential consumers demanding their products or to other firms producing their intermediate inputs.

Endogenous Growth Theory: Endogenous growth theory views innovation as the primary driver of economic growth. On the contrary to physical capital, knowledge does not face diminishing returns. As has been proven since the industrial revolution, adding more knowledge will always increase growth at the almost same rate, while adding physical capital faces diminishing returns.

Supply Chain unbundling: Baldwin also introduces a theoretical model on what globalization does to the different elements of a firm’s production chain and their respective geographic locations. His idealized version of production chains distinguishes between tasks- “the full list of what must get done to make the product”, occupations- the list of tasks performed by individual workers; and stages- groups of occupations related to each other. A central challenge in managing production processes is how to deal with the coordination-specialization trade-off. Based on this framework, Baldwin argues that the massive strides in information and coordination technology since the 1980s have had a double-edged impact on production chains: on the one hand, ICT makes the coordination of various tasks easier, thus favoring worker to specialize on fewer tasks. On the other hand, ICT also makes it easier for one and the same worker to take over more tasks. In sum, the number of tasks per occupation may decrease or increase as a result of the ICT revolution. In picking geographic locations in turn, firms try to meet a trade-off between production and separation costs. Offshoring obviously reduces firm’s production costs, but it also creates separation costs: goods need to be transported, ideas transmitted and people need to travel between factories. A typical equilibrium would involve production stages requiring high skill labour being bundled in “headquarter economies”, say in the United States, and production stage requiring a lot of low wage labour being brought together in factory economies, say in Mexico. Another factor influencing firms is that coordination costs are at their highest when half a company’s production stages are offshored. Companies often delay initial offshoring decisions, but then suddenly offshore most production stages.

In Chapter 7, Baldwin ties together the story of the two unbundlings, using the theories he previously laid out. During the first globalization, which started in the late 18^th century, once growth had taken off in the Northern hemisphere, circular causality reinforced industrialization and growth in the Northern hemisphere, while the global South, to include ancient civilizations like India and China, specialized in non-industrial goods. As long as the movement of ideas remained costly, innovation-driven (“endogenous”) growth was mostly limited to the Northern hemisphere. Urbanization was also a phenomenon mostly found in the Northern hemisphere, as cities represented a good way to economize on communication costs. On the contrary, the second globalization led to Southern industrialization, or at least industrialization in those countries which participated in global production chains. The rapid decline in communication costs weakened previously existent agglomeration forces. At the same time, some countries in the global South with sufficient natural resources to sell, such as Brazil or parts of sub-Saharan Africa, did not get to industrialize but still profited from a hike in global commodity prices.

In the final part of the book, Baldwin assesses the implications of his argument about the two globalizations for developing and developed countries. A central tenet of his is that too much economic thinking and policy advice is still based on economic theories developed based on the first unbundling. Baldwin, instead, aims to offer policy advice for a world in which production processes have increasingly become unbundled.

For economically advanced nations, Baldwin recommends governments to focus their competitiveness policy on promoting production factors which combine low international mobility, so that a country can have an actual advantage from investing in that factor, and a positive spill-over, so that an increase in the production factor translates in higher economic growth. Such production factors include highly-skilled labour, tacit knowledge which is difficult to codify, and social capital. Industrial policy should no longer focus on promoting manufacturing per se, but rather manufacturing-related service jobs which provide the highest added value. Also, governments should promote cities, as they are the right place for innovation driven by the clustering of high-skill labor. Rather than attempting to preserve specific jobs, social policy should concentrate on helping individual workers adjust. As for trade policy, the emphasis will need to shift from helping companies sell goods abroad to supporting global value chains. Possible measure include protecting companies’ intellectual property and supporting the connectivity of production facilities spread over different countries.

In Chapter 9, Baldwin discusses how developing nations can adjust their policies to the second unbundling. Traditional development policies aimed at a “big push” in order to move countries out of an agricultural equilibrium into an industrial one. Often with the help of import substitution policies, nations then aimed to climb up value chain by developing whole industrial sectors, from at first manufacturing non-durable consumer goods (e.g. clothes), to durable and intermediate goods, (e.g. cars) and (e.g. car engines). Whole regions like East Asia could be climbing up this value chain together, a phenomenon described by the so-called flying geese paradigm. On the contrary, development policy after the second unbundling should no longer prescribe that a developing country would have to promote whole industrial sectors. What counts instead is participation in global production chains. Rather than championing a national car industries, countries could instead focus on participating in certain, particularly labour-intensive stages of car production. Developing countries could enhance their competiveness by protecting the intellectual property of potential investors and by assuring that cross-border flows between countries can be kept open.

In his concluding thoughts, Baldwin outlines factors which could decisively change the world economy in coming years: a change in trade costs, for instance through rising protectionism; further changes in the costs of moving ideas enabled by ICT; and a change in the costs of moving people. A most likely scenario for a third globalization would be if the costs for face-to-face interaction, traditionally high due to high costs for moving people, plummeted massively. Telepresencetechnology could become a key enabler for even further separating the separation of production processes.

Overall, an absolutely fascinating read. Easy to read, the book still helped me better understand the development of the world economy, particularly since the 1980s. The only minor criticism is possibly that Baldwin’s model of globalization may be overtly focused on technology. While difficult to assess in its relative importance, politics may be as much of a factor in determining the costs of the movement of people, goods and ideas. The pax Brittanica may have had an important effect on the first globalization in the 18^th and 19^th century; the rise of populist protectionism in the Trump era may be massively harmful to multinational companies operating global production chains.  

A scandal and a classic

Vladimir Nabokov: Lolita

A literary masterpiece on the difficult topic of child abuse.


Looking back at his life from a prison cell, Humbert Humbert tells the story of his relationship with the young girl Lolita. Humbert grew up in Europe sometime in the early 20th century. His first teenage love, Annabelle, dies of typhus, leaving Humbert with a life-long obsession with pre-pubescent girls or "nymphets" as he refers to them. In an attempt to settle down and break with his obsession, Humbert weds a Polish woman in Paris, a marriage which ultimately ends in divorce. After his divorce, Humbert moves to the United States to take up a heritage. Working as a private  teacher, Humbert encounters 11 years old Dolores, or Lolita as Humbert calls her. He decides to give in to the advances of Lolita's mother, Charlotte Haze, in order to stay close to the daughter.
Charlotte finds him out after a short time, but then abruptly dies in a traffic accident. Humbert begins  a restless road trip throughout the United States with Lolita. Lolita is showing herself increasingly reluctant vis-à-vis Humbert's advances, but Humbert manages to subdue her through a mixture of inducements and threats. Finally, Lolita manages to run off with another elder man who claims to be her uncle.
Having searched without results for two years, Humbert finally receives a letter from Lolita. It turns out Lolita is pregnant, about to get married to a war veteran and needs financial support. The elder man she had originally left Humbert for, Clare Quilty, had been an acquaintance of Lolita's mother. At first in love with him, Lolita left Quilty when he tried to have hear feature in a pornographic movie. Realizing that Lolita won't return to him, Humbert heads off, kills Quilty, and has himself captured by police.

A masterpiece. The novel still reads as scandalous as it probably did at its publication. From a normative perspective, it is difficult to bear a story of child abuse told from the point of view of the perpetrator, who only occasionally considers the victim.  Still, Nabokov is a powerful enough narrator to leave one not only with a feeling of disgust, but also one of pity. What Nabokov offers is first and foremost a story of a man who seems to be slowly losing his mind and whose life unravels step by step. For me, it remained unclear whether Humbert was in the end still a man who could be held accountable for his deeds or not.
Nabokov also makes masterful use of the English language. The ironic distance he attributes to Humbert, the sophistacted language in which Humbert recounts his life story, and also the episodes of lucidity, remorse and regret makes Humbert an almost likable character. Humbert's slow descent into a frantic mania is impressively described. This is one of the few examples of a well-narrated stream of consciousness, a narrative technique I have otherwise never been a fan of.

Favorite quotes:
"It occurred to me that regular hours, home-cooked meals, all the conventions of marriage, the prophylactic routine of its bedroom activities and, who knows, the eventual flowering of certain moral values , of certain spiritual substitutes, might help me, if not to purge myself of my degrading and dangerous desires, at least to keep them under pacific control."

"As I look back at those seasick murals, at that strange and monstrous moment, I can only explain my behavior then by the mechanism of that dream vacuum wherein revolves a deranged mind; but at the time, it all seemed quite simple and inevitable to me.

"At the hotel we had separate rooms, but in the middle of the night she came sobbing into mine, and we made it up very gently. You see, she had absolutely nowhere else to go."

"And I catch myself thinking today that our long journey had only defiled with a sinuous trail of slime the lovely, trustful, dreamy enormous country that by then, in retrospect, was no more than a collection of dog-eared maps, ruined tour books, old tires, and her sobs in the night, every night- the moment I feigned sleep."

"Alas, I was unable to transcend the simple human fact that whatever spiritual solace I might find, whatever lithophanic eternities might be provided for me, nothing could make my Lolita forget the foul lust I had inflicted upon her."

"It had become gradually clear to my conventional Lolita during our singular and bestial cohabitation that even the most miserable of family lives was better than the parody of incest, which, in the long-run, I could offer the waif."

Very Best Friends

Philip Roessler: Ethnic Politics and State Power in Africa: The Logic of the Coup- Civil War Trap

A great contribution to the literature on organized violence, coups and civil wars.

A "coup civil-war trap" stands at the core of Roessler's account of war and the state in sub-Saharan Africa. African leaders tend to exclude ethnic groups other than their own from political power . This increases the risk of civil war, as politically excluded groups have an incentive to take up arms against the state. So why do political leaders exclude ethnic groups ? The theory of the security dilemma provides the answer. Leaders of ethnic groups cannot commit themselves to not exploit their access to the state apparatus and security forces to overthrow the sitting ruler. Incumbent leaders, in turn, cannot credibly promise to not exclude political rivals and representatives of other ethnic groups from their coalition. Under these circumstances, mutual mistrust my spiral in an ethnically inclusive ruling coalition. Coups and purges offer an attractive tool for preventive and preemptive "first strikes", permanently eliminating political rivals by restricting their access to the state apparatus and security forces.
Roessler supplements this model of ethnic politics with a theory of what he calls cooperative counterinsurgency. Most African states lack a strong state apparatus. The only possibility for counterinsurgent forces to thus collect information and apply targeted reprisals  are patron-client relationships embedded within ethnic networks. Put differently, the best counterinsurgent forces are to be found within the ethnic group whose very members are staging the insurgency. Those networks can facilitate a bargain that solves a prisoners' dilemma between local communities and the state: the state promises not to resort to indiscriminate violence, while local communities commit themselves to not support rebel groups.

Rössler deploys an impressive, multi-method approach based on country but also statistical analysis to both build and test his theories.  An extensive comparison of uprisings in Sudan's Darfur region illustrates the commitment problem of power-sharing in weak states and the cooperative counterinsurgency hypothesis. An Islamist regime, dominated by Omar al Bashir and Hassan al Turabai, had risen to power through a military coup- also labelled as the al-Ingaz (salvation) revolution- in 1989.  In the early 1990s, this regime could draw upon Islamist networks among non-Arab ethnic groups in the Darfur region, particularly among ethnic Fur and Zaghawa, to crush a burgeoning rebellion through a cooperative counterinsurgency approach.
In the course of the 1990s, however, in an illustration of the commitment problem, Bashir and Turabi entered into a power struggle, with each fearing political marginalization by the respective other. The origin of this power struggle lay in the dual power centers which the regime brought into power in 1989 had. On one hand, a Revolutionary Command Council, headed by Omar Bashir took charge of the state apparatus. At the same time, the Tanzim, the Sudanese branch of the Muslim Brotherhood, continued to exist as a shadow government. Turabai, the Tanzim's founder, took a powerful position behind the scenes.  The power struggle ended in Turabi being pushed into the opposition. The Islamist movement and its National Congress Party split along ethno-regional lines, with Turabi seeking support in Western Sudan and Bashir drawing upon the support of Northern riverain Arabs. Ethnic polarization ensued and, when another rebellion broke out in Darfour, the Sudanese state apparatus could no longer rely on Islamist networks among the Non-Arab tribes to organize a cooperative approach to counterinsurgency. Bashir had to resort to indiscriminate violence, and mostly relied on ethnic Arab militias in its counterinsurgency campaign in Darfour.

The case study evidence is accompanied by a statistical analysis of countries in Sub-Saharan Africa between 1946 and 2005, based on the Ethnic Power Relations (EPR) data-set. The EPR constitutes a large-scale attempt to code ethnic politics, to include the degree of exclusion of ethnic groups. In a first statistical analysis (Chapter 8), Roessler demonstrates that politically excluded ethnic groups are, compared to their included counter-parts,significantly  more likely to implicated in a rebellion against the state, but less likely to engage in a coup d'état.

Roessler also investigates the risks of coups and rebellions in two types of regimes in which power-sharing is borne out of necessity and commitment problems are ripe: co-conspirator regimes and split domination regimes. 
Co-conspirator regimes tend to be prone to collapse, when allies which came to power together fall out and wage war. Roessler discusses the examples of Chad, Liberia and Congo.

• In Chad, Hissène Habré, an ethnic Gouran, came to power in a rebellion during which he aligned with ethnic Zaghawa, to include the man who would ultimately depose Habré, Idris Déby. In the late 1980s, mutual mistrust rose between Habré and various Zaghawa leaders, who feared being excluded from the regime. After a failed coup in 1989, Déby fled to Sudan, from where he organized an ultimately successful rebellion.
• A similar story unfolded in Liberia in the 1980s. Two childhood friends, Samuel Doe, of the Khran ethnic group, and Thomas Quiwonkpa, of the Gio ethnic group, seized power in a coup in 1980. Doe became head of state and Quiwonkpa commander of the armed forces. In a spiral of mutual mistrust, Doe ultimately deposed Quiwonkpa, who fled the country and went on to found the National Patriotic Front of Liberia. This group staged a rebellion from neighboring Sierra Leone and, under the leadership of the infamous Charles Taylor after Quiwonkpa's killing, ultimately manage to capture power in the 1990s.
• The Congo wars in the 1990s also illustrate the risks of co-conspirators falling out and tensions erupting into civil war. Laurent Kabila took power with the help of Rwandan sponsors and, at first, incorporated a high number of Rwandan representatives and ethnic Tutsi into his government. Fearing, though, that his former patrons might attempt to replace him, Kabila ended up purging his ancient allies in a pre-emptive strike in July 1998, asking all Rwandan soldiers to leave the country. Within a week, a rebellion supported by Rwanda began in Eastern Congo, triggering a war which would ultimately encompass the entire region.

In "split domination regimes" a former colonial power had usually distributed control of the security forces and the state apparatus between ethnic groups. Under such conditions, the commitment problem may become especially severe, leading to either a coup trap or ethnic exclusion and civil war.  Burundi is a classical example, where at independence ethnic Hutu and Tutsi were represented in both politics and armed forces. A failed Hutu-led military coup in 1965 led to the exclusion of ethnic Hutu from power, ultimately resulting in a decades-long civil war.

In a final extension of his theory (Chapter 10), Roessler asks what might facilitate self-enforcing power-sharing between ethnic groups. He argues that ethnic groups might agree to share power and hence accept coup risk for a lowered risk of civil war when a "balance of threat" exists. Political exclusion makes less sense when various ethnic groups have a good chance of capturing the capital by force due to their demographic strength and the geographic location of their homelands. Roessler points to Benin and Ghana as paradigmatic cases of self-enforcing power-sharing. Both countries are relatively small and divided between different ethno-regional blocs. After independence, both witnessed a cycle of coups and counter-coups, but none of the two witnessed permanent or larger-scale exclusion of any of the ethnic groupings. Roessler also employs the EPR data to provide further support for his hypothesis on how a balance of threat capabilities can induce self-enforcing power-sharing.

Overall, an impressive book. The wealth of the empirical material, to include the results of very substantial field research in Congo and Sudan, is nothing less than breathtaking. Roessler does a great job in translating the bargaining theory of war, mostly developed to explain interstate relations, into the domain of ethnic politics. While the book limits itself to Sub-Saharan Africa, the basic logic of how power-sharing induces commitment problems should travel well to other cases as well. Nikolaos van Dam's book about how the Syrian regime came to be dominated by a small clique of ethnic Alawi around Hafiz al-Assad comes to mind. The only pitfall I see in the book is that it could have profited from some more editorial work. The presentation of the argument appeared slightly convoluted. Roessler introduces various, admittedly noteworthy, extensions to his theory in different parts of the book, making it difficult to follow up on all of them.