|
Sub-title is Understanding Iraq, Afghanistan and Vietnam - Casement
Publishers, Oxford, England, 2015, 360 pgs., bibliography, notes,
illustrations, graphs, tables, appendices
|
|
| |
Reviewer's Comments - As the author describes it in chapter 25, this is
two books in one. The first (and great majority of it) is a description of his
research and analysis at the Dupuy Institute, mostly under government contract,
on various factors that may be dominant variables influencing the outcome of
'insurgencies' (which side will win). And the second is a brief but strong
advocacy essay urging that much more effort be devoted to the subject by
funding more similar research and analysis. The subject is the analysis of
factors favoring the 'insurgents' versus the established regime in many
conflicts since World War II. The description proceeds in general chronological
order reporting on Dupuy Institute work starting in 2004 when they were tasked
with estimating outcomes for the then current war in Iraq. In a lengthy book
format it proceeds through discussions of their further work and chapters
devoted to assessments published by other authors and organizations.
The 25 chapter titles and 10 appendices well describe the subject of each. Some
state the conclusion of the analysis, such as: "Chapter 4 - Force Ratios
Really Do Matter" and "Chapter 5, Causes Really are Important".
The method is to collect a very large volume of data on a wide variety of
'insurgent' conflicts and subject it to statistical methods and models that
represent the interaction of various factors. Since some of the same data was
used as the basis for reports over several years and on related specific
issues, there is some redundancy between chapters.
This is an important book for several reasons: 1 - the results of the analysis.
2 - the disbelief by official DOD. 3 - the failure of US government to
understand from the begining what would happen. 4 - a case study of the
research and analysis methods used. 5. It is very relevant to the current
military/political policies still being followed today.
|
|
| |
Introduction -
Lawrence explains the early Dupuy Institute concern over conditions in Iraq in
2003-2004 and the initial contract to study what the future would portend.
|
|
| |
Chapter I - The Iraq Casualty Estimate
- This establishes the briefing that resulted from the initial contract,
delivered in Dec 2004 and January 2005, as the base for the subsequent
chapters. In this report the Dupuy Institute predicted that the continued
conflict in Iraq was "a major insurgency" that would last for 10
years and result in very significant casualties. The chapter provides a
description of the origin of the contract, the methods used and results
presented to military officials. Lawrence writes that the results were not
believed. There had not been sufficient time (only a few months) for full
research and analysis of the large data base the Institute already possessed to
provide more elaborate substantiating evidence. But one of their conclusions
was that the insurgency already had 20,000 to 50,000 men in the field, vice the
5,000 the government was claiming. This initial analysis was based on a data
set of only 28 historical examples.
|
|
| |
Chapter 2 - The Art and Science of Counterinsurgency Warfare Studies
- This chapter is a general historical background discussion of the broad
subject of 'insurgency war' and the thoughts about it by various military
authors. Lawrence quotes Clausewitz on conventional war by big armies and
writes that Clausewitz did not focus on the several insurgencies that took
place during the Napoleonic wars; such as in Spain, Austria and Russia. He
considered that such guerrilla war was a subsidiary front in a general war. But
Clausewitz did specify 5 conditions necessary for an insurgency to be
effective. But even then support as part of a regular force was necessary.
Lawrence discusses Charles Caldwell's book "Small Wars' devoted to the
colonial wars of the 19th century. He continues through the series of reports
and manuals published by the U. S. Marine Corps based on experiences in the
Caribbean, Mexico, and Central America. He mentions Major Earl H. Ellis's study
and those of Majors Samuel Herrington and C. J. Miller, and others.
|
|
| |
Chapter 3 - The Acid Test: Predicting the Present
- In this chapter Lawrence presents some of the analysis and predictions made
by the Dupuy Institute (for instance, First Gulf War, Bosnia, and Iraq and that
of several other well-known authors.
|
|
| |
Chapter 4 - Force Ratios Really Do Matter
- In this chapter Lawrence states the principal conclusion of the entire book.
Of all the factors analyzed and described in the following chapters none is as
critical to outcomes as the real force ratio between the incumbent government
and the insurgency. Determination of what that is can be rather slippery. Which
components of the government count and how does one determine what is the
actual number of effective insurgents?
For the Dupuy analysis they returned to a larger data base of 83 cases.
Lawrence provides 5 pages of tables. He stresses that the Dupuy conclusion on
the critical nature of force ratios was not the common belief.
|
|
| |
Chapter 5 -Cause Really is Important
- The author presents another 3 pages of data and tables to support his
conclusion that the second most significant factor is the nature of and
strength of belief in a 'cause', especially nationalism or religious.
|
|
| |
Chapter 6 - The Two Together Seem Really Important
- Here the author combines these two variables with 7 pages of data to show
their synergy.
|
|
| |
Chapter 7 - Other Similar Work
- Lawrence here digresses into discussion of the 'fallacies' of competing
'experts'. He takes on a fairly extensive list of popular authors.
|
|
| |
Chapter 8 - Outside Support and Structure of Insurgencies
- The 83 cases provide 5 pages of case data and tables
|
|
| |
Chapter 9 - Rules of Engagement and Measurements of Brutality
- These are two separate but similar variable factors. The rules of engagement
consider especially the use of artillery and aerial bombing and their impact on
the general population. The 'Brutality' relates to the use of torture, murder,
police methods, and the like. There are 14 pages focused on the effort to
determine if these are significant as operational issues. The categories are
broken down into details on, for instance, 4 specific rules, 2 on the use of
torture, 3 on controlled versus extensive use of firepower, 4 of brutal counter
insurgency and insurgent methods. Lawrence divides the characteristics into 6
categories of use by incumbent versus insurgent. And then he subjects these to
statistical analysis. he uses ratios of 'civilian" deaths to insurgent
deaths.
But I question just who should be counter as a 'civilian' when the insurgents
are integrated into the 'civilian' population and not some group out in the
jungle..
|
|
| |
Chapter 10 - Sanctuaries, Border Barriers and Population Resettlement
- These also are separate but similar variables. Their analysis required
considerable data and thought. Lawrence writes that he could not find a clear
correlation between the degree of outside support and outcomes. But sanctuaries
are important. They existed in only 21 of the 83 cases and were significant for
providing great assistance to insurgents. He did a statistical analysis using
Fischer's exact test on border barriers used by the regime in 13 cases and
found that the French border system in Algeria was very successful - but at
great expense. The border barrier created by Morocco versus the Polisario was
also successful - but in desert area.
Extensive effort at population control -regrouping people occurred in 10 cases.
The French effort in Algeria failed.
His general conclusion on the impact of 'outside support' was that he found
little correlation between this and outcomes.
|
|
| |
Chapter 11 - Estimating Insurgence Force Size
- In this chapter the subject shifts to the issue of availability of data; both
the need to be able to estimate the actual size of the insurgent force and the
need to determine the nature of insurgency in terms of motivation. He produces
on direct data estimate.
|
|
| |
Chapter 12 - The Value of Elections
- There are several categories or types of the government type involved in
conducting elections ( are they really free elections?) These range from
democracies to powerful dictatorships. The regime won more often when there
were elections and when no outside help was involved. Democracies also did
better.
|
|
| |
Chapter 13 -The Influence of Terrain on Insurgencies
- Lawrence created 12 different types of terrain -such as percent urban,
covered (meaning jungle); rough (meaning hills or mountains etc.) And he
divided categories by the difference in result between those with foreign
intervention and pure domestic cases. He found that foreign assistance did help
with it provided technical means to overcome problems with terrain.
|
|
| |
Chapter 14 - Other Issues
- These include duration of the war, type of war, type of insurgent, and
wounded versus killed ratios. Statistical methods are used in an effort to find
possible correlations of these variables with outcomes.
|
|
| |
Chapter 15 - The Burden of War
- This is a strange one - Lawrence measures the 'commitment' and 'intensity' of
the intervention by outside power in terms of percent of personnel committed
per 100,000 of the country's population - and the intensity in terms of numbers
killed per 100,000 of national population.
I do not understand the relevance. Seems to me 'commitment' might be measured
in terms of the size of the foreign force being sent as a percent of the total
military establishment of the foreign country - For many that is much more
important than the total population, which may be of little consequence to the
leaders sending their forces. Or it may be measured in financial and economic
terms in terms of the cost of the commitment relative to the country's
financial strength. Lawrence establishes 4 categories to measure commitment and
conducts statistical tests for data bases of 83, 62, 36, and 26 cases. He does
similar analysis in terms of 'intensity' for intervening forces, government
forces and insurgent force versus civilian population sizes.
|
|
| |
Chapter 16 - A Model of Insurgencies - In this chapter Lawrence
combines the data and analysis described in the previous chapters to develop a
'model' - a two variable 'model' which is depicted in graphical format.
I am not clear on what he means by 'model'.
But the graph is constructed with vertical axis showing the estimated
probability of blue ( regime) success and the horizontal axis depicting the
force ratio. Then there are three curves shown to depict three levels of
'political' concept or motivation of the insurgent side. 1 - 'central idea
" meaning the insurgents have strong belief in a concept such as
nationalism communism, religion. 2 a 'medium' commitment to an idea and 3 a
purely local or factional motivation such as local disputes or personal
motives. The three curves then show that the greater the trend to number 1, a
strong ideological commitment, the larger the force ratio favoring the regime
will be required for success.
This chapter contains the main predictive point of the study and book.
|
|
| |
Chapter 17 - Other Theorists
- Lawrence compares his ideas with those of 11 well-known authors who have
established reputations as theorists on insurgency. These are Manwaring,
Trinquier, Joes, Galula, Clutterback, Fall, Kitson, O'Neill, BDM report, Lutwak
and Peters. He goes into considerable detail to describe the main views of each
of these folks on whichever of the variables they have discussed from among
those Dupuy Institute has considered. Most of these experts only discussed some
of the many variables considered in the book. Lawrence then presents the
comparisons in tables - for instance one that has columns for Rules of
Engagement - Use of Torture - and Use of Firepower. A larger table has columns
for Area, Population, Population density, Border length, Percent arable land,
Percent urban terrain, Terrain, and Location. He groups these authors into
'schools' such as British and French opinion. Lawrence does well for the reader
by identifying what the main factors influencing results of insurgency war were
postulated by each author. And also he tried to find what each author wrote
about each of the other factors they considered. He found that most authors
ignored various of the possible factors.
|
|
| |
Chapter 18 -The Other Side
- Lawrence in this chapter summarizes the published opinions of 8 authors who
were significant leaders in a well known insurgency.
|
|
| |
Chapter 19 - Withdrawal and War Termination
- Lawrence writes that the Dupuy Institute raised this subject with officials
but not generate any interest or a contract to investigate. he considers this
an important issue. How does it all end ultimately, in compromise or
disintegration. The official 'end' of an insurgency may not be the final end.
He notes cases in which the 'loosing' insurgent nevertheless came to power
years later.
|
|
| |
Chapter 20 - Relating a Force Ratio Model to Iraq
- This chapter amounts to a critique of the American performance and actual
outcome of the counter-insurgent phase of war in Iraq. Here Lawrence cycles
back to the initial chapter and issue. He notes that the U. S. strategy in Iraq
varied over time. In particular, the strategy changed in 2005 (after the first
Dupuy report) from a 'stay in bases' concept which failed to a 'return to
combat' featuring a 'surge' in 2006 that continued to the end of 2009. He again
discusses the initial Dupuy estimate presented in 2004 with comment on each
point in the Power Point briefing slides. These include issues such as the
border, rules of engagement, treating of population and political compromise.
Then comes the fundamental conclusion. The U.S. paid the Sunni tribes into
supporting the U. S. war and fighting the AQ. The U.S. bought off the
insurgents and enrolled 100,000 of the potential insurgent fighters into its
side. In terms of force ratio the Dupuy model for 2006 showed the U.S. had a
4.8 to 1 advantage for a 86% chance of win - and after that change it had a 14
to 1 ratio with a 99% chance of win. The 'official' plan ( he disputes this was
a real plan) was for the 'surge' of American troops but had no mention of
buying out the insurgents. He writes that with the payment to the insurgents
changing the balance the long term result of that alone would have been the
same with out any 'surge' of American troops. Moreover, the concept and
proposal for recruiting the Sunni did not come from above but from local
American commanders. We bought off the insurgents rather than attack them.. But
in 2010 there still were 20,000 insurgents versus the 100,000 now 'friendly'
Sunni tribal locals. The result was that when we stopped paying the Sunni
tribes, in 2014 there was a new insurgency. On pages 246-247 Lawrence presents
actual force sizes with head counts for Sunni tribes and adds totals to show
the real force ratios. On page 249 he shows that the situation turned around
with change in force ratio.
|
|
| |
Chapter 21 - Relating a Force Ratio Model to Afghanistan
- Lawrence describes the initial U.S. overthrow of the Taliban government in
2001 - ironically an example of a successful insurgency aided by outside U.S.
forces. He notes that the Dupuy Institute did no work on Afghanistan as the
reversed insurgency grew between 2002 and 2011. But he contends that the
official U. S. estimate for size of the insurgent force (manpower) was
consistently too small. He gives an estimate of 15,000 to 25,000 versus the
official number, 11,000. This translates into a force ratio of 7.6- 10.4 to 1.
His 'regression' model then predicts a 90% chance for an insurgent victory. He
shows the same 'model' graph again with the same two questions - size of
insurgent force and its motivating beliefs - is it a regional or only factional
insurgency? Lawrence discusses a briefing presented in December of 2008 that
asked what the US views on the nature of the insurgent force really was since
it did not seem to him the US was winning. In 2009 he wrote another warning but
it was not published.
So, he asks, 5-years later what is the situation? On page 263 he returns to the
'buy in' of Sunni in Iraq and cites the statistics, 5 lessons and observations.
1 The US Government did not control - 2 it was a mistake to focus on building
police versus real military - 3 the contrast between small 'surges' and buying
off the enemy - 4 the rules of engagement - and 5 Did the US lose due to under
commitment in 2001-2004. The same mistakes as in Iraq are being made in
Afghanistan. What is the political will? What about considering the losses as a
% of home country population. What is causal relationship?
|
|
| |
Chapter 22 - Relating a Force Ratio Model to Vietnam
- Lawrence again applies his force ratio graph "model", this time to
Vietnam. He claims the US did not 'win'. Then he asks the question, "how
could the US have won? What actual force ratio would have brought a great
chance of victory? This he will study with use of the 'model' that depicts
force ratios and 'commitment' to a cause. On page 275 he poses the question,.
"Was the insurgency broadly based?" Was it wrapped up in nationalism.
He calls the insurgent army strength 200,000 to 300,000.
I have to dispute some of his estimates for size and composition of the
'insurgent forces' since they were largely as 'outside' forces and were the US
and allied forces. He includes the so-called 'TET' offensive on 1968 - I was
there. On pagfe 276 he claims there was 'no plan' and misquotes General Palmer.
|
|
| |
Chapter 23 - Conclusions
- In this chapter Lawrence summarizes his conclusions from use of his 'model'
the force ratio and commitment graphic. He repeats his conclusion that force
ratios and insurgent causes are important variables on which to build such a
'model' and that his explains 80% of the outcomes of the 83 cases. His
quantitative analysis is based on the largest such data base of 83 cases.
|
|
| |
Chapter 24 - Where Do We Go From Here?
- Lawrence writes that the work described in the book is only a good start
toward full analysis of the factors influence insurgent wins and is based on
Dupuy Institute work only through early 2008, since they have had no funding
since then. And moreover, there is little academic work on these issues. He
writes that the government lacks an attention span and only thinks of the most
immediate short term crisis. There is a lack of interest in long term studies.
He is pitching for more funding.
He presents several recommendations. The Dept. of Defense needs three sets of
quantitative predictive tools. 1 - a model that predicts the chance of
political violence wherever it might arise. 2 - a model the predicts the
chances of insurgencies in time before they occur. 3 a model of the nature of
insurgencies in being already.
I always question the concept that the future can be predicted through
computer-assisted 'modeling'.
Then the DOD needs training tools including 1. in the political concept of
motivation and cause of insurgency. 2 structure of insurgencies. 3 the question
of the role and value of outside support for insurgencies
The U. S. so far has negative learning and failed to understand the Iraq war
due to emphasis on Viet Nam.
Then we need to study the changes in violence caused by climate change.
This last one seems to me to be an obvious effort to appeal to current
government politics.
|
|
| |
Chapter 25 - A Tale of Two Books
- In this chapter he presents his full pitch, mentioned in my initial comments.
He writes that the book was conceived as a rep;ort on the theoretical base of
Dupuy Institute studies and analysis done for DOD. Having written it, he now
also has the idea for a second book (included in summary here). This is less a
theoretical study and more of an advocacy for work on current issues. He
believes the U.S. is not studying guerrilla warfare enough although now locked
in such war in Iraq and Afghanistan. Lack of study and understanding has led to
failure to create better plans or execute what ideas they have had. His pitch
is that these are American Wars of the American people. One cannot blame
Presidents Bush or Obama or Congress. Blame for failure lies with the citizens
- taxpayers and voters who are responsible for who runs the government. The
second book is intended to generate reforms to help educate everyone to the
need for better government. There is a need for long-range analysis programs to
understand warfare.
|
|
| |
Appendix I - Briefing Slides from January 2005
- Here is prints the 13 Power Point view graphs containing the results and
recommendations of the Dupuy study in 2004-5. He wants to prove the analysis
offered then was correct, which it was.
|
|
| |
Appendix II - The Bosnia Casualty Estimate
- This is a summary of the report Dupuy Institute prepared in 1995 for the
government under contract to estimate the potential casualties that would
result from the Army participating in 'peace keeping' operations in Bosnia. The
Institute had 3 weeks to provide its answer. The basis for analysis was a data
base of 144 contingencies previously created by Dupuy in 1985 plus new date
provided by UN offices on peacekeeping. The resulting estimate was for very
minimal casualties, much lower than was mentioned by others.
|
|
| |
Appendix III - List of Cases
- This is a chronological list of the 83 cases used in the Dupuy study plus an
additional list of 27 cases now available for study.
|
|
| |
Appendix IV - Force Ratios
- This table lists the cases in order from the lowest force ratio of .38 in the
peacekeeping operation in Libveria 1990 in which the insurgents won to the
highest force ratio of 162.73 in Cyprus in 1955-59 in which the interveening
force won.
I have to question when I see a number based on very gross raw data carried out
to two decimil places.
|
|
| |
Appendix V - Force Ratios as Divided by Political Concept
- Here the same cases are listed by categories of 'political concept'. First
list is for limited that is regional or factional based insurgencies, again
beginning with Liberia 1990. Then there is a list of insurgencies called
'central idea' like nationalism starting with Indonesia in 1945. And then comes
a list of insurgencies based on 'overarching' idea like communism starting with
UN PK in Cambodia 1991. Finally there are a few cases not classified. The text
discusses conclusions from these tables in terms of percentages of 'wins' for
regime or insurrection in the different types of insurgent 'causes'. One such
conclusion is: When a force is facing an insurgency based upon a central or
ovearching idea, that force will loose the insurgency if they do not have at
least a 5-to-1 force ratio. This is based upon 14 cases.
|
|
| |
Appendix VI - Results of Testing the Model Back to Data
- This table has columns with the defining one a list of estimated probability
of 'blue' victory from .0579 to 1.00 and the other columns are 'force ratios',
'true outcome', - 'concept', and 'predicted outcome'. The purpose is to test
how often the true outcome matches the predicted outcome which was based on the
combination of force ratio and concept.
|
|
| |
Appendix VII- Characteristics of Selected Modern COIN Barriers
- The barriers are classified as: Static elements - active elements -
|
|
| |
Appendix VIII - List of all 83 Cases by Indigenous Government Type
- In this table the cases are shown in columns titled: number - name - type -
election - duration in years - winner - and operation. In a separate table
those same cases in which there was support for the regime by inervening powers
are again listed.
|
|
| |
Appendix IX - Staying the Course: (an analysis of duration of
insurgencies)
- In this section Lawrence discusses wins and losses in categories of how long
the insurgent war lasted. There are quite a few small tables and much
commentary on them. His discussion revolves around his concept that level of
commitment can have as its proxy an estimate of the number of troops committed
in relation to the size of the interveening country's population. I simply do
not understand why such a correlation can be presumed. The correlations all
involve use of 'Fischer's exact test" with resulting numbers carried out
to 4 decimal places. Again, I find this an example of over-exactness in
relation to the input data.
|
|
| |
Appendix X - Data on 62 Insurgencies used for the Test of Anthony James
Joe's Theory
- This table is provided as a reference to the discussion of Joes theory.
|
|
