7.5x

... r J 5.23;.

In ‘*

" w‘ 42: .—

"4‘. _V< A
'f. Egan.

. "a.

:5;
'13"

 

 

‘ x' a}!
‘2‘;-

4‘ s
v

4 Wk 32W 2%“

. x
" 4% mg???
m r 4 -
gg'eﬁMt’xﬁf

'1! .1. Q‘ 4

.113
I - | .
"fr”ii‘ “

.l
V

* ‘ ~12-
13m

~32 L 1‘
"1:67 ‘1'},‘1'5Vﬁ: ’
3

on Q15
. ‘5'“0” 15$"? Ii
“ "évﬁvhﬁ’f if" '1‘
Ii'f'ﬁ'f Lt V' [d ’
‘31“ ’ '

n ‘u
311% 13‘9“! ‘
'4 - "57'3“
1'26 4;
'v‘

9;!
. \55 ”it“?
._v 4" ‘V:

j: fl“. ,'
x4 ' ‘ $24"
A.-

h
“£52
"f. @316“ ”33 Q.
~ 1: . “,1 .
53%;. 16 '7‘ u
v- . I L ‘ l"
M- ’f "5

‘2’

|
, 1|‘y
."N'H

“ 9
- A J:
é Ir: 3"?

, ,5’E‘58.

:3“ '. . '-

.f-f 1 . “32:257er
s.

v . 5

L m?"-

‘ '1‘»!

ﬁgﬁﬁW '
‘ 554?

V

1-,.

o
.J

I

‘ .
.33;

4'

f.

.
W

I l‘
,‘I

"-
I;
‘7

 

 

 

\iiiiiiiiii ii i; ii iii

,5 ‘_ LIBRARY
( " “1, Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

Explanations and Predictions of Political

Success of the Heads of Government since 1945

presented by

Meeyoung Ju

has been accepted towards fulﬁllment
of the requirements for

Ph.D degree in Political Science

Major professor

 

 

Date November 21 ,1994

MS U is an Afﬁrmatiw Action/Equal Opportunity Institution 0-12771

 

-———r~ - _
‘ -h
m-

me: It RETURN BOX to man than chockout from your Me
To AVOID FINES return on or baton data duo.

DATE DUE DATE DUE DATE DUE

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU to An Mandi“ Action/Equal Opportunity Institmion
amps-m

EXPLANATIONS AND PREDICTIONS OF POLITICAL SUCCESS
OF THE HEADS OF GOVERNMENT SINCE 1945

By

Meeyoung In

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Political Science

1 994

real power in decision-making. Mostly, they were the presidents, the prime ministers,
and the general secretaries over 137 countries from 1945 to 1991 including leaders who
were in ofﬁce in 1991. A two-equation model, which is composed of a probit selection
equation as well as an ordered probit substantive equation, is used to estimate its
parameters. The results suggest that the model has not only explanatory but also
predictive value.

In comparison of the reduced models (i.e., social background model, career
background model, and institutional attributes model), the results reveal that the career
background model best explains political survival. However, the institutional attributes
model accounts for the strength of survival. Nonetheless, the overall goodness-of-ﬁt
of the interaction model is better than others. And, the regional models contribute
different explanations to political survival. Finally, the descriptive analysis of personal
backgrounds shows the similarities and differences of leaders using the variables that
can not be included in the empirical models because of high percentages of missing

data.

To my mother and late father,
With Love and Gratitude

iv

ACKNOWLEDGEMENTS

I wish to thank all my committee members, Brian Silver, Nick van de Walle,
Larry Heimann, and Jerry Weinberger, for their guidance, advice, and support. I also
would like to thank family members who have been very caring, and supportive without
losing their patience. Without friends, my days in America could have not been easier
and happier. To all of my friends, whom I have known and met in America, I must

express a real big "Thanksl".

TABLE OF CONTENTS

INTRODUCTION ............................................. 1
CHAPTER I Review of the Literature on Political Leadership ............ 8
Classical Views on Political Leadership .................. 8
Empirical Views on Political Leadership ................. 10
A. Great Man Theory .............................. 13
B. Social Background Theory ........................ 15
C. Trait and Personality Theory ....................... 17
D. Situation Theory ............................... 20
E. Interaction Theory .............................. 23
CHAPTER 11 Theoretical Perspectives of Political Success .............. 29
A. The Concept of Political Success .................... 33
B. The Concept of Head of Government ................. 39
C. Theoretical Model .............................. 45
CHAPTER III Theoretical Propositions ............................ 52
. A. Do Social Backgrounds of Leaders Really Matter for Their
Political Survival? .............................. 56
B. Do Career Backgrounds of Leaders Have Inﬂuence on Their
Political Survival? .............................. 58
C. How Do the Institutional Attributes Inﬂuence Political
Survival? .................................... 63
CHAPTER IV Operationalization and Methodology .................... 76
A. Data Collection and Operationalization ................ 76
Data Collection ................................ 77
Operationalization .............................. 82
B. Methodology .................................. 85
Statistical Analysis .............................. 85
Descriptive Analysis ............................ 96
CHAPTER V Presentation and Discussion of Results .................. 99
A. Statistical Analyses of the Two—Equation Model ........ 101
Analysis of Social Backgrounds ................... 101

vi

Analysis of Career Backgrounds ................... 104

Analysis of Institutional Attributes .................. 107

Analysis of Interaction Model ..................... 114

B. Analysis of Regional Models ..................... 123

CHAPTER VI Descriptive Analysis of Personal Backgrounds ............ 133
A. Comparisons by Region ......................... 135

B. Comparisons by Duration ........................ 145

C. Comparisons by Type of Head of Government ......... 154

D. Comparisons by Number of Entry .................. 159

CHAPTER VII CONCLUSION ................................. 163
APPENDICES ....................... ' ....................... 178
APPENDIX A. Data Sources and Data Codebook ......... 180

APPENDIX B. Hazard Rate (Life Tables) .............. 188

APPENDIX C. Ordinary Least Squares Estimation ........ 189

LIST OF REFERENCES ...................................... 190

vii

LIST OF TABLES

Table l Dummy Variables ..................................... 90
Table 2 Coefﬁcient Estimates for Social Background Variables .......... 103
Table 3 Coefﬁcient Estimates for Career Background Variables .......... 106
Table 4 Coefﬁcient Estimates for. Institutional Attributes ............... 109
Table 5 Success of Prediction in Political Survival ................... 113
Table 6 Coefﬁcient Estimates for Interaction Model .................. 116
Table 7 Success of Prediction in Interaction Model .................. 120
Table 8 Goodness-of-F it of Selection Equations ..................... 122
Table 9 Signiﬁcant Variables in Regional Models ................... 124
Table 10 Goodness-of-F it of Regional Models ....................... 130

Table 11 Homogeneity Scores in Social Backgrounds, Career Backgrounds,
and Skills and Personality By Region ...................... 136

Table 12 Homogeneity Scores in Social Backgrounds, Career Backgrounds,
and Skills and Personality By Duration ..................... 146

Table 13 Homogeneity Scores in Social Backgrounds, Career Backgrounds,
and Skills and Personality By Type of Head of Government and
Number of Entry .................................... 156

Table 14 Ordinary Least Squares Estimation ........................ 189

viii

Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6

Figure 7

Figure 8
Figure 9
Figure 10
Figure 11

Figure 12

Figure 13
Figure 14
Figure 15
Figure 16

Figure 17

LIST OF FIGURES

Political Success .................................... 51
Distribution of Duration (Percentage) ...................... 78
Cases Excluded from the Sample ......................... 78
Age Groups by Region ............................... 137
Educational Level by Region ........................... 137
Majors by Region .................................. 140
Foreign, Military, Prison/Exile, and P01. Activity Experiences

by Region ....................................... 140
Political Experiences by Region ......................... 142
Political Skills by Region ............................. 142
Age Groups by Duration .............................. 147
Entry Mode by Duration .............................. 147
Foreign, Military, Prison/Exile, and P01. Activity Experiences

by Duration ....................................... 149
Political Experiences by Duration ...................... . 149
Political Skills by Duration . . .. ......................... 151
Party System by Duration ............................. 151
Age Groups by Type of Head of Government ............... 155
Political Skills by Type of Head of Government ............. 157

ix

Figure 18

Figure 19
Figure 20
Figure 21

Figure 22

Foreign, Military, Prison/Exile, and P01. Activity Experiences

by Number of Entry ................................. 159
Political Experiences by Number of Entry .................. 160
Personality by Number of Entry ........................ 160
Hazard Rate of Total Sample (N=984) .................... 188
Hazard Rate of Censored Sample (N=882) ................. 188

INTRODUCTION

Research on political leadership has drawn upon psychological, sociocultural,
historical, politicoeconomic, and philosophical concerns. Even though it has been
discussed with interrelated conceptual themes (i.e., leaders, heroes, elites, exercise of
inﬂuence, power, authority and control, legitimacy, etc.) and with a multidisciplinary
perspective, leadership has eluded any easy and simple theoretical explanations.
Assessing and measuring political leadership is very complex and difﬁcult since almost
any propositions concerning it present a series of contradictions, tensions, and
antinomies. An enormous variety of new conceptual deﬁnitions, contending theories,
and incongruent standards of measurement appears to have complicated rather than
simpliﬁed the comparative analysis of political leadership (Edinger, 1975). Whether
the study of political leadership is productive depends on the deﬁnition of political
leadership and the models which the investigators choose to employ.

Political leadership is a timeless and universal phenomenon, and all societies at
all times have their leaders. Political leaders come in many forms, with many styles
and diverse qualities, and at different times. There are some quiet leaders and some
famous or notorious leaders one can hear about even in neighboring countries. Some
ﬁnd their strength in eloquence, some in judgment, and some in courage. Leaders seem

to be good or bad, heroes or villains; but leaders seem also to be more or less

2

successful, more or less effective (Blondel, 1980, p.277).

When we say ’here is an effective leader, a wise statesman, a just ruler’, or, on
the other hand, call a leader a tyrant, an opportunist, or a fraud, we make comparisons
between what we see in a leader and what we expect of one (Edinger, 1976).
Deﬁnitely, there is a need for a systematic and empirical analysis of the achievements
of leaders. For example, we can ﬁnd an analysis that so far has only been in a partial
manner about some revolutionary leaders (Rajai and Phillips, 1979 & 1983).

When we evaluate the ﬁtness of candidates for certain positions, judge the
performance of political leaders, or compare our political leaders with those who aspire
to displace them or rule other people, we usually compare and generalize the
characteristics or the issues that they hold and the situation in which they are found.
It will be worth comparing present political leaders against the leaders of bygone days,
between different countries at different time periods, and generalizing about the features
that make them successful or unsuccessful. Although psychological aspects of political
success are difﬁcult to analyze and are just beginning to be examined systematically in
the context of individual leaders, it is possible to frnd the broader social backgrounds
and career proﬁles of political leaders, political and social environments, as well as the
characteristics of the rise or fall of leaders that have been easily documented.

In a simple and highly abstract model, a set of factors determines an outcome
in terms of winning-losing or success-failure. Questions of leader’s success and failure
are both empirically and nonnatively important; yet scholarly writing in that area suffers
from a lack of clear criteria (Sinclair, 1993). Political leadership of society, whether

democratic, autocratic, or monarchic, at least initially derives its power and authority

3

from the support or acquiescence of one or more dominant segments of society. If
political success is related to the concept of ’power’, we can say ’tenure is power’.
Some political leaders are more successful than others, while some are not successful
at all. Typologizing leaders on the theoretically important dimensions of social
backgrounds, career backgrounds, personality, and environmental differences, would
permit explanation and prediction of leaders’ performance in ofﬁce on the basis of their
durations of ofﬁce.

Assuredly, the best leaders - perhaps the only good ones - are those who view
their task as that of a moment, although perhaps a major moment, in their country’s
history (Blondel, 1987, p.47). However, in this dissertation, the main focus is not on
whether a leader is good or bad, but on what makes him successful or unsuccessful
based on his duration ‘in ofﬁce. Although there have been several arguments for
explaining political success in different ways, political success based on a leader’s
performance might be indicated by how long the leaders stay in offices as a head of the
government.

In this study, political success is viewed as political survival. The longer
political leaders stay in their ofﬁces, the more successful they are, regardless of whether
they exit peacefully and constitutionally or forcedly and nonconstitutionally. Some
leaders with a long duration of ofﬁce got out of their ofﬁces in such constitutional ways

as retirement, resignation, defeat in election, or natural death.1 At the same time, some

 

’ The examples of natural death are Enver Hoxha in Albania (4] yrs), Mao Tse-Tung in China
(27 yrs), Sekou Toure in Guinea (24 yrs), Seretse Khama in Botswana (13 yrs), Gamel Nasser in Egypt
(15 yrs), Forbes Bumham in Guyana (14 yrs), Jomo Kenyatta in Kenya (14 yrs), Jawaharlal Nehru in
India (16 yrs), Francisco Franco y Bustelo in Spain (34 yrs), Sékou Toure in Guinea (24 yrs), Chang Kai-
sek in Taiwan (25 yrs) Eric Williams in Trininad and Tobago (19 yrs), etc. And, the examples of other

4

leaders who stayed in ofﬁce for a long period of time were deposed by coup d’etat,
assassinated, or expelled by force.2 In the latter case, can we consider them successful
leaders? How about the leaders with a short stay and a peaceful exit? Nonetheless,
considering that ’duration of ofﬁce’ is an element of personal position and that ’disposal
of leader’ is an indicator of the strength of personal position, the subject of political
survival is more related to the duration of ofﬁce as a general concept.

Even though it is possible to expand on several questions, this study will focus
on why some leaders stay in ofﬁce longer than others. Measuring success or failure is
not easy; however, it is possible to see whether a leader stayed long enough to achieve
satisfaction as an ofﬁce holder. Whether or not a leader advances his goals is a solid
initial basis for gauging degree of success or failure. In.order to accomplish what he
sets out to achieve, he needs to maintain his ofﬁce as long as possible. The polity is
managed for survival of the leadership rather than for equitable development of the
polity. To hold an ofﬁce longer as a top executive, political stability should be
guaranteed. Even though durability is an elusive concept since relatively frequent

changes of chief executive may conceal a good deal of continuity in the parties and

 

peaceful exit after long duration of ofﬁce are Todor Zhivkov in Bulgaria (36 yrs), Wladyslaw Gomulka
in Poland (14 yrs), Robert Menzies in Australia (16 yrs), Mathieu Kerekou in Benin (19 yrs), Ne win
in Burma (19 yrs), Konrad Adenauer in West Germany (14 yrs), Tage Erlander in Sweden (23 yrs),
Augusto Pinochet Ugarte in Chile (15 yrs), Mathieu Kerekou in Benin (19 yrs), Ahmadou Ahidjo in
Cameroon (22 yrs), Yumjagiyn Tsedenbal in Mogolia (32 yrs), Anker Jorgensen in Denmark (17 yrs),
Urho Kekkonen in Finland (16 yrs), Walter Ulbricht (21 yrs) and Erich Honecker (18 yrs) in East
Germany, etc.

2 The examples of forceful exit are Janos Kadar in Hungary (33 yrs), Ahmed Sukarno in
Indonesia (16 yrs), Norodom Sihanouk in Kampuchea (16 yrs), Moktar Ould Daddah in Mauritania (17
yrs), Alfredo Stroessner in Paraguay (35 yrs), Feminad Marcos in Philippines (20 yrs), Jaaffar el-Nemery
in Sudan (16 yrs), Leabua Jonathan in Lethoso (19 yrs), Moussa Traoré in Mali (23 yrs), Nicolae
Ceausescu in Romania (26 yrs), Siyad Barreh in Somalia (22 yrs), etc. Anastasio Somoza Garcia in
Nicaragua (19 yrs), Park Junghee in South Korea (17 yrs), Ngarta Tombala in Chad (14 yrs), Anwar el-
Sadat in Egypt (11 yrs), etc. were assassinated.

5
personnel in power, durability of political leaders can be an indicator of political

stability.

Even if we can not study political leadership and political leaders in a full sense
of the word, we can limit the scope to the people in certain positions and analyze them.
The sample for this dissertation will be the heads of national governments from the end
of World War II to the end of 1991. While ofﬁcial titles differ from country to
country, a head of government may be deﬁned as a person who really has been or is
in power in each country. Top executives are almost always interested in maintaining
oﬁices and the power of the purse has long been crucial in their search for security.
The security of top executives is generally proportional to the length of time they hold
ofﬁce; newly elected leaders are the most insecure.

From 1945 to 1991, more than 1,000 individuals have been at the top at their
countries. Of course, many others sought but did not achieve such a post. Some stayed
in office for longer periods than others and held a position more than once. This study
will include relatively large numbers of comparable cases (997 cases) which lend
themselves not only to making qualitative statements but also to quantitative statements,
statistical analyses, and empirical generalizations. Thus, ﬁnding a good method of
classifying qualitative data for this study is very important. Modeling with the variables
that imply certain characteristics, i.e., duration data, high percentage of missing data,
and several dummies, should be done with care.

This dissertation will focus on the comparative analysis of political leadership
across countries, combining the basic themes and questions that delineate different

aspects of political success or survival. There have been a range of philosophical

6

opinions on political success examined by philosophers qua philosophers, over time, and
in the West vs. the East. Philosophers mainly deal with depicting the nature of an ideal
leader. Some philosophers focus on the nature of human beings from which the
concept of political. leadership derived. These fundamental philosophical perspectives
on the idea of political leadership would provide the requisite attributes that an ideal
leader should have.

However, I Will review more recent literature on political leadership success
based on more theoretical or empirical speculation. It will explicate the images that we
really see in a leader. Most of the prevailing studies of political leadership are analyzed
in a descriptive way. Through descriptive analyses, it is possible to identify several
important factors that make leaders different or similar by comparison. Basically,
examining those factors - such as social background, career background, personality and
skills, and contexts - will allow us to do predictive and causal analyses on political
success or survival.

The main purpose of this study, however, is to determine whether the subject of
political leadership survival as indicated by political success can be studied in empirical
and comparative .ways. Furthermore, introducing a quantitative method with qualitative
data will test how personal qualities, that is, individual differences on social and career
backgrounds, and institutional differences are related to political success or survival.

Thus, this study will put emphases on models that will test: (1) What are
important factors that determine political success?; (2) ‘Do social background factors
inﬂuence political success?; (3) Does political success depend on career background

variables?; (4) Do situational or institutional factors have effects on political success?;

7
(5) What subset of factors among the above can best explain political success?; (6) If

a leader is successful, what makes him or her more successful or less successful than
others?; (7) Do qualities of political leaders change across the different regions or with
the systematic differences in politics?; and, (8) Can we expect some individuals with

some qualities would be successful political leaders in the future?

CHAPTER I

Review of the Literature on Political Leadership

In this part, I will review more recently-explored literature that focuses on
theoretical or empirical aspects of political leadership instead of any speciﬁc deep
philosophical reviews. Although the principles of proper political leadership are based
on generalized assumptions about the nature of man, society, sovereignty, and authority
emphasized by several political philosophers, those normative views focus on a question
of ’who should govern’. In order to look at ’who in fact did rule’, it is more
appropriate to focus on empirical views that discern those characteristics of politicians

that would suggest explanations and predictions of their effectiveness or success.

Classical Views on Political Leadership

Political philosophers used to be interested in ethics: What makes good
leadership - for the led as well as for the leader himself? They outlined what some
would label an ideal of leadership, although it is presented as a model to be emulated
by ordinary people functioning in the routine circumstances of daily life. Western
philosophers have never been silent on the question of leadership. From Plato’s
philosopher-king to Machiavelli’s prince, from Hobbes’ sovereignty to Nietzsche’s
Ubermensch, philosophy has said its piece on the nature of the ideal leader (Grob, 1984,
p.263).

Especially, in ’T he Prince’ and ’The Discourses’, Machiavelli devotes the core

of his analysis to the study of the way in which rulers should behave. His analysis is

9

' devoted to the elaboration of technical and tactical moves designed to ensure the
success of rulers and their maintenance or expansion in ofﬁce. The substantial purpose
of leadership is subordinated to the problems of the stability of ruler in ofﬁce since, to
be certain, stability is an overriding goal in politics.

Montesquieu had more interest in government that is sensitive to sociological
and ecological differences of every country. Those ecological conditioning factors
provided certain settings that condition political systems to be constituted properly. To
him, whether it is a natural property of small states to be governed as republics, of
middle ones to be governed by monarchs, or of large empires to be ruled by despots,
it follows that in order to preserve the principles of any established government it is
necessary to maintain the existing size of the state; and that the nature of the state will
change to the extent that the state constructs or extends its limits (Corrinthers, 1977).
Also, Saint-Simon emphasizes the men who are most ﬁtted to manage the affairs of the
nation and possess the most eminent, varied, and most positively useful ability for
managing national interests and maintaining public order.

Later, Max Weber contributed many things to the study of political leadership,
introducing three types of legitimate rule or authority. He tended to regard
nonlegitimate ’power authorities’ as unstable and short-lived. He was not much
concerned with the speciﬁc ways in which leaders achieve power, how long they are
in office or how they exit. Only to the extent that duration and exit mode may affect
the legitimacy of the whole system are these problems examined especially in the
context of charismatic authority.

Eastern philosophers also have been interested in political leadership. Not only

10
the Confucian philosophers such as Confucius and Mencius but also the Taoist, Lao

Tze, proclaimed that government ofﬁcials should be responsible for establishing virtue
throughout the realism. Confucianism urges rule through the ’power of virtue’ and
moral suasion, whereas the Taoist image of leadership indicates nonleadership such that
’the greatest conqueror wins without joining issue’ (Lao Tze, 1961). The idea of
’superior-man’ is based on the harmony of a perfect individual and a well-ordered
society to bring peace to the people.

While the Confucian tradition of rule has been regarded as the rule by men, not
by law, the legalist image of leadership by the manipulation of rewards and
punishments was introduced by Han F ei Tzu. Recently, Mao’s political moralism
(pragmatism) was rooted in ancient Chinese tradition and he, in fact, was not the ﬁrst
political thinker to view the moral reform of citizens as the province of political
leadership. On the basis of his virtuocratic and meritocratic values, the military arts
were not highly valued. Soldiers oﬁen stepped in to impose their will on a disorderly
or corrupt political scene, but no cult of the warrior ever emerged to supplant the

civilian values of Confucianism.

Empirical Views on Political Leadership

Modern developments - to the detriment of ethical and philosophical
considerations - such as various theories, research methods, and psychological
implications have had the effect of emphasizing the quasi-scientiﬁc aspect of political
leadership (Hodgkinson, 1983). The focus of interests in political leadership has been

shifted from theory to research. Over the past decades, political leadership or political

ll

elites have been considered as an important issue in the area of political science. As
Plato’s The Republic or Gorgias, Machiavelli’s The Prince, and even Confucius’ The
Analects mainly deal with what is an ideal leadership, these recorded classics could
provide some theoretical explanations.

However, we need some empirical research on action-relevant understanding,
theoretical elaboration, comparison, experimentation and prediction. Even though we
can set up an image of ideal political leadership, it is impossible to have those ideal
leaders. We just select somebody as a likely potential winner, or preferable to others
as our leaders. Nonetheless, it is not certain that leaders with ideal qualities would
survive long in a very changing situation.

Research on how political ﬁgures are perceived by the public has proceeded in
several aspects; citizens judge leaders, such as presidents and prime ministers, by the
party they represent or with which they are afﬁliated (Campbell, et al., 1960). The
public conceptions of political leaders are associated with such diverse elements as
leaders announcing positions or policies they promote or oppose (Weisberg and Rusk,
1970; Markus and Converse, 1979) or the consequences of such policies (Hibbs, et al.,
1982). Also, some researchers put emphases on personal qualities such as honesty,
intelligence, and integrity; or competence of the leaders such as past political experience
as important characters in leader and candidate experience.

More recent leadership theory tends to be regarded as a subdivision of
psychology or sociology rather than of philosophy. Burns (1979) summarized such a

trend as follows:

12

Historians and biographers focus on the unique person with more or less
idiosyncratic qualities and traits confronting particular sets of problems
and situations over time. In a meanwhile, psychologists scrutinize
generic factors, early intrafamily relationships widening arcs of personal
interaction, changing constellations of attitudes and motivations, whereas
sociologist view the developing personality as it moves through a series
of social contexts - family, school, neighborhood, workplace. Political
Scientists emphasize the social and political institutions impinging on
developing leaders, changes in political leaders as they learn from their
experience, the eventual impact of leadership on policy and on history

(p.26).

The sequence of explanatory research, for instance, has been from maxims or
rules of thumb through trait theory, to factor analytic trait theory (yielding the classical
two dimensions of task orientation and person orientation), to situational qualiﬁcations,
to interactive considerations (task plus leader, followers plus interactions), to the
latter-day reﬁned maxims of path-goal analysis, and to the current complexities of
dynamic models (Rustow, 1967; Hodgkinson, 1983). Also, political scientists have
proceeded far beyond Lasswell’s generalizations and typologies in exploring the
psychological dimensions of their subject (Rustow, 1966, p.687). A number of
biographies combined political and psychological analysis with careful attention to the
canons of historical evidence (Neustadt, 1960; Edinger, 1965).

Even though speciﬁc studies of political success have not been done by many
scholars, there has been quite a few works closely related to this theme. If a complete
answer to the question about what makes successful leadership is beyond the sc0pe of
most studies, a partial response may nevertheless be assayed. They are basically
divided into two broad categories that focus on individual political leaders and

institutional processes and outcomes. The ﬁrst category is primarily based on individual

13

qualities such as social backgrounds and career backgrounds. The second deals with
the analyses of the performance of political leaders in the aspect of political stability

and political development.

A. Great-Man Theory

For a modern analysis of individual leaders, a parent of our modern approach
to political leadership is the great-man theory discussed by Thomas Carlyle (1775-
1881). According to him, the history of the world is the Biography of Great Men. As
German historians saw in the human past the ever-changing manifestations of timeless
national genius, Carlyle’s hero worship was inﬂuenced by them. As one of the
conservative theorists, Carlyle (1966) viewed democracy as the everlasting privilege of
the foolish to be governed by the wise and to be guided in the right path by those who
know better than they do.

His great-man resembles Plato’s philosopher-king in some respects as the belief
that political rule should be exercised by the man of exceptional wisdom and ability.
However, a heroic leader is an indispensable agent of historical evolution and progress;
his distinctive quality is the capacity for bold and audacious action based upon an
initiative, rather than a philosophical grasp of historical reality. Meanwhile, his idea
is close to the spirit and tradition of Machiavelli in the sense that the hero must be
immersed in the equivocation and compromise of actual life. Consequently, he provides
a needed emphasis on the role and importance of superior individuals in politics. He
says sincerity - a deep, great, and genuine sincerity - is the ﬁrst characteristic of all men

in any way heroic.

14

Most philosophical reviews on political leadership have focused on the theme
similar to the great-man theory. However, the ’great-man theory’ can not be easily
proved in an empirical way. As an example, if Socrates’ maxim "know thyself" is
fundamental to leadership, the most important characteristic of leaders is that they must
understand themselves before they can work on others. It seems to be a test for
self-realization in leaders themselves. Whereas the ’great—man theory’ asserts that
leaders are born and not made, most empirical studies emphasize that leadership is
almost accidental, or at least that most leaders emerge from circumstances which
normally do not create them. The construction of leadership would span time, cultures,
and circumstances.

’Great men’ do make history, and the cause of real intended social change can
be tracked back to the purposes and decisions of the most visible actors on the political
stage. Carlyle’s heroes, Nietzsche’s great blond beast, Hegel’s evoker and carrier of
the spirit of the times3, and Sidney Hook’s event-making man4 are the good examples
for this theory. Basically, leadership is viewed typically in terms of the great man who
shapes for a long period the destinies of mankind.

The great-man theory has inﬂuenced some scholars (Morgenthau and. Hein,

.1983) in the ﬁeld of political leadership. The question, ’what is greatness in any man

 

3 Hegel (1952, p.295) asserted: " The great man of the age is one who can put into words the
will of his age, and tell its age what its will is, and accomplish it. What he does is the heart and essence
of his age; he actualizes his age."

" Hook (l986)’s theory on political leadership is similar to Carlyle’s great-man theory.
However, Hook named it ’the event-making man’ in history. ’The event-making man’ is an eventful man
whose actions are consequences of outstanding capacities of intelligence, will, and character rather than
of accidents of position. In a meanwhile, the eventful man in history is any man whose actions
inﬂuenced subsequent developments along a quite different course than would have been followed if
these actions had not been taken.

15

and more particularly what is greatness in a statesman?’ is one of their inquiries. The
great man is the one who at least approaches perfection in certain qualities of mind,
character, and action that illuminate the very nature of man. Those who become leaders
do so by virtue of their possession in superior degree of certain personal traits - such
as stamina, decisiveness, composure - which make it their fate to reach positions of
leadership in their organizations or societies (Tucker, 1977). According to this theory,
the characteristics of social backgrounds and attitudes might be introduced and

explained as good criteria for the great-man.

B. Social Background Theory

The major reason for focusing on the social antecedents of leaders is an implicit
assumption that if we know who they are, and were, and how they came to be leaders,
then we already have a reliable guide to what they believe and do. According to the
researchers on this theory, social origins are important, justifying their focus on arguing
that social background is a useful indicator of the characteristics of a political leader
(Crewe, 1974). Also, background data on the social and career characteristics of
political leaders provide no infallible guide to either their attitudinal patterns or policy
preferences. Biographical analysis of these characteristics can reveal much information
about the nature of the political system since hypotheses concerning the functional
relationship between the requirement of the political system and the social and career
attributes of the political leaders (Waller, 1976, p.155).

To see what kind of man becomes a political leader, it is necessary to examine

the existing theories and hypotheses conceming the social backgrounds of political

l6

leaders. Social background theory is based on the question: Does the political leader’s
socio-economic status and previous life experience make any difference in the type of
head of government, leadership characteristics (e.g., charisma or non-charisma), and the
type of recruitment mechanism? In turn, the following questions are generated: (1)
How can the study of social backgrounds and career patterns of political leaders
contribute to an understanding of why recruitment arrangements function as they do?;
(2) From what social positions are political leaders recruited?; (3) What are the
interrelationships between their characteristics and political change or revolution?; (4)
What are the effects of the characteristics of political leaders on the conduct of
government?

In fact, who becomes a political leader is not just decided by elections,
primaries, and voting but, rather, is also the result of a continual sitting and sorting of
the citizens who enter the quest for political power. Careful research and imaginative
theorizing about trends in the social backgrounds of political leaders may supply some
new insight into the phenomena of social and political change. In other words, changes
of the trends in their social backgrounds may be an indication of overall social,
economic, and political change.

The historians, biographers, and journalists who have gathered most of the data
on social backgrounds of political leaders have overlooked the universal for the unique,
the general for the speciﬁc. However, what is needed to obtain a new understanding
of political leaders is an approach that emphasizes the similarities rather than the
differences between them. As Rustow (1967, p.684) mentioned, political writers sought

to catalogue in their ’mirrors for princes’ the virtues of the ideal ruler and the principles

17

that must guide his conduct. These ideas are less normative and more empirical than
those of earlier theorists and are, in turn, the basis for more recent research in this
general tradition.

Even though some researchers reject social background analyses, others (Rustow,
1966; Edinger and Searing, 1967; Farrell, 1970; Beck et al., 1973; Camp, 1979, 1980)
are interested in a social background approach. Despite the masses of accumulated data
and radical changes in the way in which data are collected and presented, one ﬁnds
rather few general relationships between social facts and political behavior emerging
from these studies.

Rustow (1966) agreed that the study of the ruler’s social background is only a
supplement, not a substitute, for the study of their performance in ofﬁce - and of the
effect of that performance on the recruitment of future rulers. Beck et a1. (1973) and
Farrell (1970) examined personal backgrounds and educational or career characteristics
of the individual leaders in communist countries. Especially, Lane (1988), recently,
analyzed the ’circular ﬂow of power’ in USSR based on social backgrounds. Camp
(1979) tried to compare women political leaders with male political leaders using

similar background data on recruitment and career patterns.

C. Trait and Personality Theory

Over the years, a group of social psychologists has tried to ascertain by
experimental methods what are the individual traits of character associated with
leadership (Truman, 1951; Gibbs, 1954; Erickson, 1958; Verba, 1961). Harold Lasswell

was the ﬁrst political scientist to introduce psychological explanations to political

18

science. He attempted to show that there are natural leaders especially equipped by
personality or skill to perform important political roles. Political leaders are
characterized by the possession of special personalities and skills, and an individual
becomes a political leader at least in part because of certain personal qualities or traits.
The same skills and traits are required to be a leader in all situations and circumstances.
Burns (1977) backs up Lasswell’s view explaining political leadership with more
speciﬁc psychological factors, such as wants, needs, aspirations and expectations.
According to Dahl (1969), political skill is generally thought to be of critical importance
in explaining the power of different leaders. However, despite many attempts at
analysis, from Machiavelli to the present day, political skill has remained among the
more elusive aspects of power.

The trait or personality theory asserts that a prominent political ﬁgure is
analyzed in order to discover the signiﬁcance of his main political role in the context
of his whole personality system (Lasswell, 1965; Barber, 1965; George, 1968;
Greenstein, 1985). Erickson (1964) claims that the success of leaders depends on a
match between their own personal characteristics and historically conditioned
characteristics of their political followers. In a word, for this study, there are several
personality and skills factors that may account for the personal greatness of political
leaders.

Foti et al. (1982) assert that different leadership labels imply different
characteristics. Mainly based on perceptional theory (i.e., social information
processing), prototypical ratings generated through several items explain which of a

series of trait and behavioral descriptions of national political leaders would covary with

19
their effectiveness. Suedfeld and Rank (1976) showed that the political fate of

revolutionary leaders, that is, success or failure, was decided by the skill in adapting to
the information-processing demands of the environment. For an example, long-term
success for a leader would be associated with low conceptual complexity (single-
minded) during the revolutionary phase, followed by a change to high complexity
(graduated, ﬂexible, and integrated) during the poststruggle phrase.

Leadership is a traitlike quality possessed by certain individuals, arguing instead
that leadership is an attribution made by perceivers that explains outcomes or
characteristics associated with actors (Merriam, 1967; Calder, 1977; Lord & Rush,
1980; Pfeffer, 1977; Foti, Fraser and Lord, 1982). Some have advanced and developed
the concept of a particular political personality or character type in pursuit of power
over other values (Lasswell, 1948; Lane, 1961; Seligman, 1950). Political leadership
focuses on what Easton (1953) has called the authoritative allocation of values in a
community and involves both the attempt to exercise and the actual exercise of power
to make policy-choices between alternative courses of action affecting this allocation
of values.

Especially in case of the revolutionaries, most researchers emphasize
psychological explanations. Lasswell’s paradigm for power-seeking explains a
contemporary mechanism for feelings of inadequacy and low self-esteem, a private need
that is located on public objects and rationalized in terms of the public interest. Burns
(1979) focuses on the analysis of the interrelationships of persons on-the basis only of
power. Based on this theory, most research deals with political styles according to

which the leaders are categorized.

20

Although both the Freudian perspectives, which emphasize a leader as a father
ﬁgure, and the idea of the leader as a person who voices the ideological concerns of his
followers, are persuasive, there is relatively little social psychological research that
bears on either hypothesis. Childhood development, socialization, and ego psychology
are the related areas with trait theory. ’Power seeker’ is a person who pursues power
as a means of compensation against deprivation. Once in power, every leader wants
to stay in ofﬁce as long as possible, attempting recasting political institutions,
reinterpreting and expanding the functions of existing political roles or creating new
ones which ﬁt their needs, political styles and aspirations (George, 1968).

Most people intuitively expect leaders to have some personality traits or
attributes which set them apart. However, ’trait theory’ or ’personality theory’ could
be attacked for several reasons. According to Greenstein (1986, p.36), personality
characteristics are less important than their social characteristics in inﬂuencing behavior,
and individuals with varying personal characteristics will tend to behave similarly when

placed in common situations.

D. Situation Theory

The political struggle for power follows the same rules in all parts of the world.
Human desire to get power can be assumed to be ubiquitous, but it depends on
changeable environmental conditions that individuals and groups are capable of getting
power and using political institutions for their advantage. Several researchers have
contributed possibilities to produce a variety of propositions about the ways executives’

success related to government stability and policy leanings (Ames, 1987). Most

21
empirically-based studies of political development include a thesis that political stability

is associated with political development.

According to the situational approaches, leadership is highly situational and
contextual, and thus a leader is legitimated and gives direction to certain cultural values.
It applies to other topics related to political success, i.e., government stability, regime
legitimacy, modernization, or political development, cabinet durability or executive
durability (Seligman, 1956; Lipset, 1960, 1963 and 1981; Hurwitz, 1973; Schumpeter,
1975; Siedman, 1978; Powell, 1982; Dodd, 1984; Lijphal'ta 1984, 1992; Panebianco,
1988; Mainwaring, 1992). Powell (1982) analyzed political performance as an indicator
of political success based on three- dimensions: citizen electoral participation,
government stability, and political order. The performance of all political systems is
evaluated by durability, civil order, legitimacy and decision efficiency (Eckstein, 1965).

The legitimation-crisis paradigm explains that crisis stems from the state’s failure
to match its capacity to respond with the level of citizen’s demand and that citizens lose
their political conﬁdence in governmental leaders, particularly, top executives. The
greater capacity of government in the more modernized societies would enhance
executive stability and effectiveness (Ahnond and Powell, 1978). Variations in the
exercise of leadership are associated with such contextual factors as domestic
socioeconomic conditions, organizational structures, and cultural patterns, or with
differences between routine and crisis situations, stable and unstable national or
international conditions (Edinger, 1990).

Gun (1974) insists that the durability (persistence and adaptability) of political

systems depends upon the nature of their structures of political authority such as

22

typological traits, institutionalization, and internal coherence of authority structure.
Persistence means how long the system exists and adaptability refers to how well the
system responds to the need for changes. As a result, persistence is deﬁned as
longevity and this durability as a simple property represents whether a polity is stable.
Since failures of the system are the fault of incumbent power holders the situation can
be greatly improved by changing the incumbent authorities. The cure for a government
and an economy that are performing poorly is a change of leadership - brought about
by the democratic process (Lipset and Schneider, 1983, p.390).

It seems relevant to consider some important situational factors which determine
whether a person is chosen for a leadership position. Considering situational factors,
several recruitment studies have discussed career patterns, opportunity structure, and
social learning (Seligman, 1964a, 1964b, 1971, 1974; Schlesinger, 1966; Prewitt, 1970;
Wriggins, 1969; Stogdill, 1974, 1981). The behavior of political leaders will vary
considerably from situation to situation. According to Fiedler (1971),

We can now see quite readily why leadership trait will not predict

leadership effectiveness. If any one can be an effective leader in some

situations and not in others, it is obvious that a single leadership trait or

a combination of traits will not enable prediction in all situations. Such

a trait or combination of traits may well predict performance in certain

situations that can be speciﬁed, and this is what the contingencies model

[that Fiedler presents] shows (p.15).

After examining and comparing the relationship between personality and situational
explanations of political behavior, between personality and role explanations of political

behavior, and between personality and cultural explanations of political behavior, he

concluded that individuals’ social characteristics rather than their personality

23

characteristics determine their political behavior.

Even though Huntington (1975, p.251) praised Ayub Khan of Pakistan for
coming close "to ﬁlling the role of a Solon or Lycurgus or Great Legislator on the
Platonic or-Rousseauian Model," Ayub failed to develop legitimate political institutions
and to ensure meaningful participation by the people in the political process. The
result was a mass upheaval which led to the disruption of the country - political decay.
In developing or underdeveloped states, individual leaders tend to be far more legitimate
than either existing political institutions or the institutions they may choose to fashion.
However, in an ideal sense, legitimacy should reside not only in individual leaders but
also in the political institutions of the state. It is much the case in the economically

developed states that dominate the development indicators.

E. Interaction Theory

Recently, leadership has been based on the idea of a capable person being in the
right place at the right time. It emphasizes not only individual differences but also
different situational aspects. When political leadership is deﬁned as a pattern, it
becomes apparent that it is "both a function of social situation and a ﬁrnction of
personality, as well as a function of these two in interaction (Gibbs, 1947). It shows
that the qualities most valued in leaders relate to their ability to reﬂect the group norms
and aspirations, with an added aptitude for improving the group standards and
performances. This circularity in the relationships is particularly noticeable within
political leadership, and can be accepted as an hypothesis in exploring the nature of

leadership.

24

In addition, political leadership is related to personal qualities and circumstance
as causative factors of the phenomenon (Dion, 1968, p.4). It can be explained. not only
by the conjunction of the nature of the leader (trait theory) but also by the nature of the
event, problem, or task, confronting the group (situation theory). It is interesting to
examine how the combinations of environmental factors and leader’s backgrounds affect
the perceived success of the leader.

According to this theory, a leader will emerge as the result of a good ﬁt between
the personality of the individual actor and the characteristics of the setting (Kellerman,
1986; Ballard and Suedfeld, 1988). The social structure or situation variables determine
political opportunities, personality structure supplies the political motivation, and skills
determine the level of political success. As a consequence of the economic, political,
and technological changes, political leaders have changed from men with aristocratic,
preindustrial backgrounds to businessmen and their advisors and agents, who are
typically lawyers. Besides such personal attributes as the individual’s ability and his
educational and technical background, what environment factors inﬂuence who becomes
an effective or successful leader?

The interaction theory explains the question of ’becoming political leaders’ based
on situation, skill, and psychology (Rajai and Phillips, 1988, p.107). It will answer
three interrelated sets of questions: (1) Who are they?; (2) Why and how do they rise
to leadership positions?; and (3) In what ways are they similar and how do they differ?
Personality factors alone cannot determine who become a leader, especially since some
of the factors which determine who becomes a leader are quite outside the individual’s

control. Consequently, an explanatory factor that is psychological and a control factor

25

that is sociological can be conceived as two different levels of description. One might
regard an attitude as derivative of an objective position or status, or an objective
position in society as leading to psychological processes such as attitudes.

Edinger (1964) introduced three different models for explaining political success:
rational-behavior model, group-behavior model, and actor-centered personality model.
Based on the ﬁrst model, success as a leader is measured by the achievement of certain
objectively posited goals and failure by an ability to achieve them. F unctionality-
dysfunctionality is perceived by the investigator in terms of the manifest end values
which may be achieved by bargaining, persuasion, or coercion.

However, the second model explains that functionality-dysfunctionality is
evaluated according to means rather than ends. The third model, meanwhile, simply
focuses on the individual’s personality need, i.e., ego satisfaction and tension reduction,
in evaluating functionality-dysfunctionality. Edinger (1964) claims that leaders’ role
performance is evaluated by role congruence. Unless role congruence is restored by
changes in the situational determinants and changes in his role performance, it is not
easy for the people to agree to his role assignments and to accept him as their leader.
A test of successful and effective political leadership would seem to be apparent
congruence of the aspiring or actual leaders’ role perception with those of his counter-
players in a relevant role-set (Edinger, 1964, p.658).

Some scholars have quantiﬁed the characteristics of successful presidents and
presidential greatness (Wendt and Light, 1976; Sirnonton, 1981, 1984, 1986a, 1986b,
and 1987; Nice, 1984). Dean Keith Sirnonton (1984) developed an interaction theory

by integrating the great-man and situational theories. According to him, the eponymic

26

signiﬁcance or sociocultural success of a political leader is determined not only by the
impact of the situational context but also by the individual attributes such as
intelligence, morality, leadership qualities, productivity, aggressiveness, age, etc.
Nevertheless, he concludes that a key predictor of presidential greatness, duration of
time in ofﬁce, may have more situational than individual causes.

However, the length of service a particular leader may depend, partly, on
personal inﬂuence and, partly, on his ability to get himself renominated and reelected.
Length of reign is associated not only with the possession of qualities of leadership but
also with a nation’s economic, political, military and even cultural well-being. Another
example of the combination of situation and personal quality factors is Cell’s (1974)
work which has been progressive in the direction of assessing the different levels of
charisma displayed by various heads of state. Stogdill (1981) is also interested in the
attributes of the leaders and situations.

Blondel (1980), more concretely, analyzed political leadership using
personalities, careers, socioeconomic backgrounds, and political styles based on the
interactional theory. He systematically dealt with every national chief executive who
held ofﬁce from the end of World War II to the 19703, although he did so to provide
a purely descriptive explanation. Especially, as the question of succession related to the
popularity of the leader in ofﬁce and of the leader or leaders who might replace the
politician in power, regularity of succession related to the strength or pull of these
leaders on the population (p.102). He claims that the strength of a leader depends on
various characteristics of his own personal characters as well as the political system.

Recently, Bienen and Van De Walle’s (1991) approach to leadership longevity

27

describes and explains the patterns of risks for losing power that leaders face. Their
purpose of the study is to explore the probability of a leader’s loss of power over
different intervals of time without relating leadership longevity to power enhancement.
or regime stability. Nonetheless, their models focus on individual leader traits and
country traits based on the event-history approach.

The maintenance of political leadership position rests on a system of social
exchange or transaction (Burns, 1976). The structure and dynamics of the relations
between leaders and followers are the main themes in the transaction theory. According
to Hollander (1978), when leaders are effective, they give something and get something
in return. Power and conﬂict always exist in the relationship between leaders and
followers. He combines the situational approach with a social exchange component that
focuses on reciprocal inﬂuence between the leader and the followers.

Philosophical concern with conﬂict reaches back to Hobbes and even Heraclitus,
and men who spurred revolutions in Western thought, such as Machiavelli, Hegel,
Marx, Freud, and Feuerbach. The theories of Pareto, Durkheim, Max Weber and others
contain many concepts, assumptions, and hypotheses which greatly inﬂuenced later
writers who attempted to deal with conﬂict in general; and George Sirnmel carried
theories of conﬂict into 20th centuries. According to Hegel, Marx, Feuerbach, conﬂict
produces consciousness. In other words, leaders can make followers more conscious
of aspects of their identity, such as sexual, communal, ethnic, class, national, and
ideological, etc. Leadership style is regarded as the nature of the relationship between
leaders and followers. In this sense, agents from the perspective of followers would be

moral and political culture within which they operate and social formulations from

28

which they are drawn. Even though political leadership is explained as the interaction
between leaders and followers, it is not easy to examine or demonstrate this relationship

retroactively in the historical contexts with which we deal (Rajai and Phillips, 1988,

p.4).

CHAPTER II

Theoretical Perspectives of Political Success

The political leadership approach suffers from argument and confusion over key
terms, a relative dearth of testable hypotheses, a failure to clearly separate normative
from empirical theory and, not least, the lack of a ﬁrm data base in which the latter
could be solidly grounded (Mayser and Wagstaffe, 1987, p.1). The lack of conceptual
focus cannot be attributed to either philosophy or method, whether verbal or
quantitative, empirical or normative (Lindblom, 1968). The concepts and hypotheses
of the leadership perspective on one hand, and its methods and techniques on the other,
have been quite heavily criticized.

At a given point of time, each government has only one government; at the top
of the government is a supreme authority - whether individual or collective - a single
leader or a cabinet. Every political system must have its legitimate rulers. For
instance, formal-legal authority of individuals in key policymaking positions, that is,
presidents and prime ministers, depends on the question of the institutionalization of
effective executive authority. Individual leadership at the top is based principally on
the legitimate powers of the highest governmental positions. According to Max
Weber’s rules, some leaders gain legitimacy from legal-rational norms, some from
traditional norms, and others from charismatic norms. Authority and legitimacy imply
the recognition of a right to hold ofﬁce as well as the general consent to the rules of
the political system.

Legitimacy refers to the overall level of acceptance of the regime or government

29

30
throughout the polity. The legitimacy of authority in a given country has been

evaluated by the legitimacy of the rulers dependent on any particular standards of
righteousness or. performance. There are neo-utilitarian theories that make regime
legitimacy dependent on the effectiveness and efﬁciency of government in providing A
more life satisfaction for the worst part of society and, eventually, good and equal
satisfaction for everyone (Handerich, 1976).

For a leader’s authority as a top executive, his or her hold upon the follower’s
loyalty and willingness to follow, will depend in some measure upon the effectiveness
or ineffectiveness of the leadership performance from the follower’s point of view
(Tucker, 1991, pp.41-42). The effective control of political leaders requires creating
a stable political order and securing the legitimacy of political rule. It is persuasively
agreed that the great majority of any population has a strong tendency to see their head
of government as their nation, and to give legitimacy not only to the person but to the
acts of that person. The failure of political institutions to manage rapid social and
economic change is viewed as a major contributor to crisis of legitimacy and national
integration. And, if such failure happens, a leader might be vulnerable to reprimand.

Over the decades, political leadership has been explained by two different
perspectives. One is that leaders are viewed either as heroic and famous ﬁgures
(charismatic or traditional leaders). Another is that leaders are viewed as the agents
from the perspective of followers (legal-rational leaders). While the former is explained
by the traits, personality and skills theory, the latter is supported by the situational
theory. A precise assessment on the impact of leaders can be made in two crucial

respects; the popularity of leaders among the population and the attitudes of citizens to

31

various issues.

However, the recent trend seems to be a combination of the two perspectives.
The qualities of political leaders, the immediate moral and political culture within which
they operate, and the social formulations from which they are drawn are all parts of the
equation. It is clear that the advantage of any particular qualities of individual leaders
or situational inﬂuence varies not only from country to country, but also from time to
time in any given country. As long as the exercise of power includes a relationship
between those who exercise it and those who are subject to it, longevity has often been
linked to stability and performance (Black, 1966; Rustow, 1967; Huntington, 1968;
Eckstein, 1971).

The stability of the state and the regime and the effectiveness of the public
authorities seem to depend on the way in which the institution of the chief executive
is organized and ﬁtted with the other institutions of the nation. The longevity of a chief
executive is related to the legitimacy of the institution, partly because this will give
stability to the institution and partly because the efﬁciency of the government is also
at stake. The efﬁcient institution of the chief executive will become more legitimate
as time passes, while one which is inefﬁcient will lose credibility. The ’credibility’ of
leaders is an essential aspect of governmental life. However, the support of a leader
and his policies does not always include the gaining of legitimacy to the larger systems
of the state and the lack of support for a speciﬁc leader or a policy does not always
imply a lack of overall legitimacy.

No society remains unchanged throughout all time. As long as the problem of

political order or stability is compounded by the problem of legitimacy, analysis of

32

political success should be focused on leadership performance. Moreover, the
comparative approach calls for a common theoretical framework by means of which the
relationship between the legitimacy aspect of leadership and the socioeconomic and
political structures of several countries may be analyzed. The interaction between the
approaches to legitimacy of the leaders and of the structures within which they led can
be studied comparatively in order to ascertain whether or not there is a relationship
between this interaction and their success or failure in maintaining legitimate authority.
The nature and interrelationships of patterns of authority, levels of modernization,
methods of citizen mobilization, and levels of industrialization, etc. should be
considered as important factors in a political leadership equation.

The concept of legitimacy is elusive and it is not easy to operationalize.
However, much literature has looked for a number of speciﬁc ingredients which could
enter into the composition of legitimacy and which could be more easily operationalized
(Rogowski, 1974, pp126-27). Legitimacy and political integration are affected by
variations in general social and economic conditions. If a society is static and relatively
insulated from the inﬂuence of other polities, the relative strength of the regime or
government will not tend to change. Where integration is high, it will remain high.
Where groups and political system are not in harmony, conﬂict will remain high, as in
plural societies.

Political instability represents a crisis for a nation and its rulers. According to
Morrison and Stevenson (1971, p.348), it induces a situation in which institutionalized
authority patterns in a nation break down and political violence replaces compliance

with political authorities. Thus, political stability deﬁned as the endurance and

33

persistence of regimes over time is measured by the length of time that particular
constitutional arrangements have been in place or that modernizing leadership has been

in control.

A. The Concept of Political Survival as Political Success

In a general sense, success is explained by the idea that individual decisiveness
should be added to luck. Whereas luck is the probability of getting what you want
without trying, success is how often one gets what one wants if one tries. However,
in political leadership, much of what makes for success is out of the individual’s hands.
According to Lasswell and Lerner (1965), men seek to maximize ’valued outcomes’ by
adopting ’institutional practices’ that also affect the ’resource environment.’ One of the
valued outcomes is ’power’ outcome that includes a ﬁnal decision as winning or losing
an election or a war. Sinclair (1993) also claims that a solid initial basis for gauging
degree of success or failure is whether or not a president advances his goals. According
to one standard, if a president accomplishes what he sets out to achieve, or if he attains
his own goals whatever they happen to be, he is a success.

What constitutes a ’successful leader,’ is considered a matter of deﬁnition in
terms of the problem under investigation. He may be a man who gets things done, such
as furthering economic development or winning elections or wars; he may be a leader
who maintains the ’stability’ of a political system in terms of stress or who introduces
major innovations into political processes (Edinger, 1969, p.10). It is clearly in accord
once with the fact that successful leaders are career leaders who devote long years of

careful grooming to the political roles they eventually come to assume, whereas

34

unsuccessful leaders are crisis leaders who are susceptible to the forefront without much
preparation and typically in response to situations of national emergency. According
to Fred Riggs’ (1981) argument, the political leaders who fall into the successful
category found their legitimacy on a mixture of norms congruent with the structures
within which they led.’ As a result, their success is determined by the interaction of
their leadership strategies, the cultural symbols that they are able to manipulate, and the
institutional and social contexts within which they operate.

Support relates to legitimacy, and legitimacy stems from individual support.
Where positive and passive support exceeds negative reactions, the regime can be said
to be relatively legitimate; the greater the difference in favor of positive support, the
higher the amount of legitimacy. This enables the public authorities to continue in
office. The. citizens in a constitutional state are free to criticize openly the nature of
governmental decisions, to propose alternatives and to take forms of constitutionally
recognized action to change laws and policies. Therefore, respect for political ofﬁce
and a disposition to obey must be the primary elements in political authority (Carter,
1979, p. 33). Support is explained by two issues: (1) Who wins and who looses in
political competition?; and (2) Why and how do politics persist or change? (Eckstein,
1979)

Succession process is one useful way of observing the legitimacy of a system.

How is political leadership transferred from one man to another and from one party to

 

5 His model is called as prismatic model that combined three norms from Max Weber such as
charismatic norms, traditional norms, and legal-rational norms. If a president is a head of the party as
well as a head of government, it is not a prismatic model. The criteria to evaluate the leaders successful
or not are based on the combination of three norms.

35

another? The succession process is a useful checkpoint for looking at the question of
legitimacy because when power is transferred, individuals within the system are forced
to decide whether their loyalties are conﬁned to those who have exercised authority or
to the system of government itself. For example, the executives’ inability to maintain
consistent support has directly caused regime instability and encouraged military coups.
In the case of persistent incongruence, the voters will not support the central actor, a
revolution will overthrow him, his formal subordinates will fail to execute his orders
(Edinger, 1964, p.659).

‘ The hurt of withdrawal affects every leader, no matter what its cause and no
matter how power passes. It does not matter if a powerholder is thrown out of office,
comes to the end of a legally limited term, or leaves of his own accord; there is a
feeling of emptiness and a sense of having done something wrong. Even those who
seek power to accomplish a goal - and accomplish it - even they leave ofﬁce with
regret, a fear of boredom, and a vague sense of shame. The loss of power is the
realization of how badly they had exercised it. So from a psychological aspect, the
method of losing power would make a difference.

A successﬁrl leader holds a position as a chief executive until natural death or
the end of a constitutional term, followed by voluntary retirement. If the nature of
leadership is considered as a structured pattern of authority and subordination, the
leadership is not simply a function of the leader acting by himself, but the result of an
interaction with the people. A successful leader may be an individual whose role
interpretations are accepted by the salient counter-players because he satisﬁes their

expectations, or he may be a person who so clearly comprehends the political context

36

that he can adjust his behavior to meet the expectations of these relevant others
(Edinger, 1964, p.665).

Political legitimacy is a key to creating a viable polity capable of sustaining
national development and without it, development efforts are likely to be futile and their
achievements in administration or any other ﬁeld are apt to be short-lived. The
credibility of leaders and of institutions is an essential aspect of governmental life. If
people withdraw their consent from any leader, that leader could ﬁnd it difﬁcult, almost
impossible, to continue in a leadership role. In the end, people remove and replace any
leaders from whom they have withdrawn consent. The lack of legitimacy leads to the
lack of authority on the part of the political leadership and the lack of self-compliance
on the part of subject people. Nevertheless, in reality, the polity is managed for
survival of the leadership rather than for equitable development of the polity.

In many societies, it seems that leadership and the use of power are inextricably
woven together. Many leaders can mobilize the high legal and administrative
bureaucracies of the state in massive displays of power. However, the effective use of
power requires social legitinrization. In all sociocultural systems, leaders should
mobilize human, material, and symbolic resources of society toward speciﬁc social
ends. They should have abilities to maintain their regimes. Legitimacy is threatened
not so much by one who would break the rules as by one who seems to break trust in
the image of the presidents or the prime ministers as digniﬁed, plain and clean in
character. A system which failed miserably would lack legitimacy and would,
therefore, be a prime candidate for extinction or fundamental system change. If a

nation’s political performance, legitimacy, and persistence are all threatened, political

37

instability accrues.

Leadership success is increasingly evaluated on the basis of the fulﬁllment of
directives. Gupta (1990) stresses two capacities to gain legitimacy among the people:
the democratic form of government and the maintenance of economic prosperity. Even
though executive durability is deﬁned as the government stability in democratic regimes
in most studies, especially in parliamentary government, it is possible to apply for all
different government styles including communist systems. According to Lipset (1959),
legitimacy involves the capacity of a political system to engender and maintain the
belief that the existing political institutions are the most appropriate or proper ones for
the society. Besides the promise of economic prosperity, leaders can hold on to power
through an appeal to the national ideology.

Focused on the chief executive as a target, a measure for the evaluation of a
leader’s performance should be related to government stability or to the durability of
a chief executive. The deﬁnition of stability demonstrates considerable staying power
and a capacity to endure without great or frequent changes in pattern. All countries
emphasize on economic growth and political stability, and these are main factors to
prolong the leader’s tenure. However, it is not right to say that a deﬁnition of stability
is merely longevity, even if we would agree on just how long a democracy must endure
to be considered stable. For example, the presence of short-lived governments is taken
as an evidence of poor performance in all regimes. Also, the tenure of an emergent
political leader is frequently short-lived. Political success can be explained by stability
of political government that are measured by the tenure of a chief executive. The

primary criterion for maintaining leadership position is the satisfaction of the members

38
of the society.

The power motive is one of the most relevant explanatories to political behavior.
The desire for ego enhancement and for manipulating others have been suggested as
important in understanding political orientations. All leaders seek power, and, if they
get into power, they want to keep power as long as possible. Power as a head of
government is a function of the leader’s will to arouse and tap the needs and wants of
followers and his capacity to mobilize resources to meet those needs and wants, thereby
contriving to retain followers’ support and to continue in power (Burns, 1978, p.386).
The exercise of leadership depends more or less on his performance in the roles that an
individual in a high political ofﬁce may play.

According to Sirnonton (1986), one of the indicators of presidential success or
greatness is ’years in ofﬁce’, a variable that is used to predict most of the variance in
presidential success. It is actually based on eponym theory such that long-lived rulers
may provide history with eponyms, that is, with convenient names for the epochs of
history. In a narrow sense, a climb up the political ladder from lower status to national
ofﬁce means political success to political ﬁgures. If they become heads of government,
there is no place to move up. Therefore, in a broad sense, political success means
strong durability in ofﬁce. Political success, that is, continued tenure in ofﬁce, depends
on the impression that the ofﬁce-holder is handling his job competently.

Power comes with ofﬁce, and leadership rests on force or on consent. It is
important to combine power with authority to be an effective powerholder. Power is
often treated as synonymous with ’authority’, and authority is commonly referred to as

’legitimate’ power. To the leaders who ﬁnish their career or their life, leadership means

39

the culmination of their career. Of those who come into power, some will leave it
honored with the glow of a job well done and a memory of themselves carved into
history. Some will leave it with a feeling of frustration and failure that can reach to the
deepest levels of despair. Some will fail in the eyes of others but ﬁnd a way to look
upon it in another light. They succeed in not being successfully challenged before the
end of their working life. If they do not succeed in staying at the head of their country
until retirement, they are less successful or less fulﬁlled than those previously
illustrated. Whether a leader is formally appointed, formally elected for a speciﬁed

term of ofﬁce, he has to rely on the acceptance of his acceptance by his followers.

B. The Concept of Head of Government

While Tucker (1981) argues that leadership approach to the nature of politics is
distinct from power approach since leadership is a value-neutral phenomenon, Janda ,
(1972) asserts that a positional view relates to the concept of ’power’. It is true that
leadership is not identical to power. However, it is a certain kind of power, or it is the
use of power in some circumstances. The individuals in political institutions and groups
who exercise more inﬂuence than others on the allocation of values are considered key
concepts in explaining political leadership (Huntington, 1971, p.316). According to
Jennings (1960), political leadership is concentrated on formal leaders - the holders of
ofﬁcial positions of authority, and especially those in national politics.

Political leadership, as concept and category of analysis, is constant, while its
actual membership or personnel composition is variable. It varies from society to

society, depending upon such factors as historical tradition, past political experience,

40

foreign inﬂuence, ideological predilection of society, or simple happenstance. There
exist differences in forms of society-wide political leaders from a cross-cultural,
historical, and evolutionary perspective. The leadership of a society is a criterion of the
values by which that society lives. Also, by learning the nature of the leaders, we learn
much about the nature of the society. Political leadership is a role that is understood
in terms of the social and cultural context within which it is embedded and which
shapes the particular forms it takes in any society.

Many political scientists also deﬁne the concept of political leader in many
different ways. A leader is the one who occupies the position or ﬁlls the role that
allows him or her to wield the greatest power. Dahl’s professional politician, Snyder’s
political decision-maker, Lasswell’s policy scientists and Deutsch’s cybernetic
leadership are other examples. In addition, more speciﬁcally, Ilchman and Uphoff’s
(1971) statesman is a political actor as a head of regime. Apparently, identical or
similar conceptualizations are often translated into very different operational criteria for
empirical research; and seemingly different ones are operationalized in much the same
way (Edinger, 1969). The analyses in this study would be largely conﬁned to the heads
of government considering both positional and power-related concepts. Success of
political leaders will be regarded as one of the outcomes of political leadership, since
it depends on how they, as heads of government are evaluated or perceived by the
people in their society.

Whatever the political system is, i.e., the Western democracies or the communist
states, it is true that an individual or the collective occupies the unique position of being

at the apex of the pyramid of often absolute power. The main mediator between

41

constitutive system and bureaucracy is the head; the head of the government; the head
of state; the head of political party; the head of the armed force. Sometimes these roles
are united in one person, but often they are separated so that a polity has several heads
at once. In constitutional regimes, the head has a formal title, but under naked rule, the
most important head could be a boss who rules from behind the scenes, while those
with formal leadership titles are mere ’puppets’. For instance, the gradual displacement
of dynastic rule in Europe by constitutional monarchies and republics produced a
differentiation between a head of state, symbolizing the society as politically organized,
and a head of government.

Among all the heads, the role of a head of government (chief executive) is the
most important in practicing actual power, since the executive branch is both the prime
mover and the principal regulator of governmental performance. In modern
constitutional regimes, the chief executive is chosen through the constitutive system.
The chief executive can be chosen by the elected assembly in a parliamentary system,
by the electorate in a presidential system; or by the chairman of the party, etc. The
very nature of elections - the typical process for ﬁlling political ofﬁce in a democratic
society - entails the selection of those with good qualiﬁcations for ofﬁce. Electing
someone need not imply any belief in his inherent superiority, but it does usually imply
belief in competence and so the right, which stems from being qualiﬁed, of the speciﬁc
individual to hold that ofﬁce (Carter, 1979, p.34). Therefore, this dissertation will
include the political leaders who are or were in actual power in each regime, that is, the
heads of governments across the nations.

Regimes have frequently been classiﬁed by their most ’visible features’, which

42

has meant a classiﬁcation based on either the different forms or the different styles that
political leadership has taken. However, traditional classiﬁcation of regime type or
forms of government is more or less meaningless. Even though the division based on
the idea of a separation of powers, i.e., executive, legislative and judicial, applies to the
presidential system, it does not ﬁt into the parliamentary system.

Every country has someone at its head, but there are considerable variations in
the extent to who is at the head. Historically, executives formed the central core or
kernel of governments, out of which, separate legislative and judicial institutions
progressively emerged and broke off. For an example, the term ’executive’ may refer
to what in Britain is usually termed ’the government’ and in the USA. ’the
administration’. Or it may refer to an even narrower set of ruling institutions and
individuals who possess ultimate decision-making power, such as the Politburo and
General Secretary in the USSR and other communist states, the ruling juntas in
military regimes, and, of course, presidents, prime ministers and their cabinets
(Charlton, 1986, p.18). Some countries have a group of people running the countries,
such as a presidium, a revolutionary council, or even a council of ministers.

A classiﬁcation of contemporary types of political executives may most logically
begin by focusing on heads of governments. According to Blondel (1984), they would
be the ’varied and diverse’ nature of executive types among contemporary regimes.
They would be included in three dominant categories - monarchs, presidents, and prime
ministers. The special categories of members of ruling councils and Communist party
secretaries in communist states have to be added to those. It focuses on the leaders

who are really in charge of the overall aspects of policy making or the overall view of

43

the affairs of the government. From this point of view, all formal ’heads of state’ who
merely symbolize the nation with pure monarchy system should be eliminated. Those
nations are Afghanistan before 1973, Bhutan, Brunei, Bahrain, Kuwait, Morocco, Saudi
Arabia, Swaziland, Oman, Jordan, Qatar, and the United Arab Emirate. Also, the
presidents of the collective council (Uruguay during 1955 - 1965) or rotating presidents
(Switzland and Yugoslavia since 1980) are eliminating from the sample.

Since most contemporary state are republics, it is not surprising to ﬁnd that the
largest groups of contemporary heads of government are presidents. The limited
presidency is the most common form of political executive in the world today, about
52 countries having adopted it, embracing 25% of the global population. Presidential
systems based on liberal-democratic ideology, imply rigid forms of succession. Those
systems lead to ﬁxed terms of ofﬁce and organize a separation of powers. It does not
allow a president to stay in ofﬁce, even if he is popular, after a stated period because
he is not normally allowed to succeed himself or to stay in ofﬁce more than a limited
time.

Some presidents are elected for a limited period only and may not be reelected
immediately or, at most, may be reelected only once.6 Bans or restrictions on
reelection are for a check on presidential power as well as a safeguard against an

executive’s self-perpetuation in ofﬁce (Shugart and Carey, 1992, p.88). As the forms

 

6 In some countries such as Brazil (since 1988), Ecuador (since 1979), Guatemala (since 1986),
Honduras (since 1985), South Korea (since 1988), Mauritania (since 1980), Mexico, and Uruguay,
through the denial of immediate reelection, a time limit is put into presidential term. Where reelection
is permitted, it is allowed only after one or two intervening terms in Bolivia, Colombia, Costa Rica, El
Salvador, and Venezuela. Also, some countries, such as Austria, Liberia, Nicaragua, Paraguay, Peru,
Sierra Leone, Taiwan, Tanzania, and U.S.A., allow just second consecutive terms. However, these
constitutional restriction will be ignored for the test.

44

of authoritarian presidency, military rulers often introduce some forms of civilian
structure in order to prolong their tenure. The form of ruling collectivities or councils
is common in military regimes where the designation of one ofﬁcer as chief executive
frequently obscures the genuinely shared character of the executive leadership group
which takes power following a coup d’etat.

Secondly, historically, the prime ministers who emerged from beneath the wings
of monarchs, have entrusted originally with the subordinate tasks of leading the state
bureaucracy and coordinating the institutions of governmental decision-making. In a
multiparty parliamentary system, it can consist of a coalition of leaders of several
parties. Prime ministers emerged as the heads of the bureaucracy of the State, even
though it may not always be true to say that they are always ’ﬁrst among equals’. They
control their party and the parliamentary majority and, if electoral support is maintained,
exercise control for long periods.

The ﬁnal type of contemporary political leaders are ﬁrst or general secretaries
of ruling Communist parties. Since communist states are led by a Communist party
which is viewed as the vanguard of the nation, the leader of the Communist Party -
nominally the ﬁrst or general secretary - is in effect the country’s leader. The Party
Secretariat has, as one of its many executive functions, the duty to oversee and
supervise the ’state’ bureaucracy which is formally headed by a prime minister or a
president in the communist countries. He embodies the permanent legitimacy of the
nation. The communist leader is chosen from among existing leaders, normally
concentrated in the politburo and the Secretariat such as the hierarchical structures in

which the decisions of higher bodies places major decision-making powers in the

45

topmost organs of the Party. As far as authority or legitimacy implies recognition of
a right to hold ofﬁce and implies general consent to the rules of the political system,
in a situation, in which a leader is the main beneﬁciary of popular authority in the
nation, this leader is at liberty to organize his government as he chooses (Blondel, 1982,

p.27)

C. Theoretical Model

We generally, speak of a government as an institutionalized pattern of rule. The
rules or governs may constitute a vast and complicated design, as in modern
constitutional regimes, or they may be monocratically arraigned in one simple hierarchy,
as in aristocratic regimes, including totalitarian dictatorship. Every political system may
contain some institutionalized means of recruiting leaders. Also, every political system
may have overall ethical evaluation of the quality of their rule.

The normative question ’who should rule?’ indicates ’on the basis of what
principle of selectivity should political leaders be chosen?’ However, for the empirical
aspect, it is important to examine the leaders who in fact did rule. As far as the
concept of political leadership is narrowed to ’heads of government’ or ’chief
executives’, the political leadership can be explained in a certain view. When we
evaluate some leaders successful or unsuccessful, we can focus on how long a leader
has been in his/her ofﬁce. In general, longer the leader stays, the more successful
his/her rule is.

As discussed in the literature review, the interaction theory has dealt with the

combination of traits and situation, social backgrounds and situation, or traits and social

46

backgrounds. However, theoretical development for the analysis, here, employs a
combination of the social background theory and the situation theory except the
personality and skills theory. We can cast the search in the light of the interaction
theory of political leadership stressing the interplay of two approaches described above.
The scientiﬁc development of political leadership studies - ranging from initial
conceptualization through creation, Operationalization, and testing of hypotheses to
rigorous theoretical integration at the highest reaches of mathematical imagination - will
be promoted by pursuit of a multimodal research strategy. Nonetheless, in the
descriptive analyses, personality and skills variables are used for comparison.

The problem of keeping a leadership position is complicated because the leader’s
authority does not derive solely from the political organization. Authority does not just
ﬂow from the top down. Rather, a primary criterion for maintaining the leadership
position is how much people are satisﬁed with their leader. Whether a leader is
formally appointed, or formally elected for a speciﬁed term of ofﬁce, he has to rely on
the acceptance of his authority by people. As a result, leaders’ success in their ofﬁces
result from the attitude of the people toward them. The leaders’ social backgrounds,
career backgrounds, and situational factors are considered as main criteria by which
people evaluate their leaders.

This study will focus on predictive and causal analyses. Basically, the analyses
are not only the speciﬁcation of leadership outcomes that are the object of prediction
but also the speciﬁcation of different backgrounds that are hypothesized to predict
leadership outcome, that is, duration in ofﬁce, and the speciﬁcation of the theoretical

linkages among independent variables and dependent leadership outcome. The social,

47

economic, and cultural characteristics of the society in which the organization is
embedded are also critical inﬂuences on the nature of leadership. As a result, the main
research is concerned with the relationship between political leaders’ success and their
social backgrounds, career backgrounds, and situational factors. Whether or not a
leader becomes successful depends on the possession or acquisition of certain key
attributes and experiences.

Here, this study tries to get some answers for the possibilities of the research on
political leadership. The variable items are adapted from several general pieces of
literature of leadership analysis. The variables to be considered are based on scholarly
consensus. The variables for which data are collected interrelate to a large extent since
demographic, ideological, attitudinal, psychological, and situational variables not only
cohere as distinct groups, but also intersect at many points. Hypotheses about political
behavior that were ’tested and proven’ valid at one moment in history need to be
rechecked constantly against the changes of time and circumstances.

The models include several different possible explanations: (1) the leader’s social
and career backgrounds; (2) the causes of stability and instability of types of
government in the contemporary world; (3) the governmental policies; and (4) the
country characteristics, etc. These anticipatory variables simply alert attention to things
that may help to describe, explain, predict, evaluate, and intervene. A theoretical model

is constructed as follows:

Political Success = f (Individual Social Backgrounds, Career Backgrounds,
Institutional Attributes).

48

However, as an indicator of political success, duration of ofﬁce, has somewhat
different roles to evaluate political leadership. ’Duration of ofﬁce’ decides whether a
leader stayed for a ’SHORT’ or a ’NOT-SHORT’ period, that is, political survival.
After that, it can examine whether he is more successful than others, that is, strength
of political survival. Inclusion of the leaders who are still in power into the analysis
will need some manipulation in order to reduce misspeciﬁcation errors. The
speciﬁcation of the model is :

Pr( Duration of Ofﬁce ) = f (Individual Social Backgrounds, Career Backgrounds,

Institutional Attributes).

Truly, one of the most difﬁcult tasks in measuring and evaluating performance
of political leaders is the task of trying to look at the elements that make up political
success. The dependent variable in the theoretical model denotes the leadership
outcome, whereas a variety of causative factors that extend from very general
explanations to highly speciﬁc reasons are regarded as the independent variables.
Various measures of selected independent variables contribute importantly to the
explanation of variance in measures of duration of ofﬁce. Independent variables are
divided into three different subsets: demographic attributes, career attributes, and
institutional attributes.

Individuals’ personal backgrounds are divided into two groups, demographic or
social attributes and career attributes, with the former being caused by nature and the
latter beeing caused by experiences. Political scientists on political leadership assume

that research on this subject should focus on the behavior or attitudes of political

49

leaders. Therefore, interviews or surveys with them would be the best way to get the
data for the research.

However, political leaders, as heads of government, especially at the international
level, are quite often inaccessible, and attitudinal information is therefore difﬁcult to
obtain. There are relatively few variables which are susceptible to quantiﬁcation that
are of greater concern to the political scientists than those describing the backgrounds
of political leaders. Even though it is possible to appreciate on a cumulative basic
changes that are occurring in the characteristics that deﬁne political leaders worldwide,
if possible, information on political leaders should be kept and analyzed as qualitative
data.

The demographic (social background) backgrounds consist of age, place of birth,
education, major, father’s occupation, and father’s SES, which are considered as family
legacy and political socialization. The career backgrounds are represented by political
ideology, party membership, experience in congress, experience as a Minister, previous
occupation, military experience, experience in prison or exile, political activity, long-
term foreign experience, enter mode, and numbers of entry as a head of government.
These variables could be the criteria for evaluating whether a leader has enough
qualiﬁcation for handling the ofﬁce as a head of government.

Personal disposition variables imply the aggregate characteristics such as
personality and skills that deﬁne a unique human individual (physiological factors;
psychological factors, etc.). Even though the role of personality appears to be large,
there still exist difﬁculties in measuring and assessing. A lot of studies of personality

have dealt with a number of elements such as energy, courage, intelligence, etc.

50

However, this study, relating to the concept of ’power’, just focuses on ’psychological
traits’ and ’task-oriented’ traits. Oral skills, organization skills, and language skills,
physiological status, loss of parents or separation from parents in early age, and power
connection are included in this subset.

Based on a theory of democratic development, institutional, socioeconomic and
cultural aspects are very important. As institutional variables, several country
characteristics and political systematic variables such as population density,
homogeneity index, freedom status, party system, economic freedom, literacy rate, type
of head of government, and region as cultural difference, and etc. are used for the
models. In an institutional aspect, the degree of individual freedom and participatory
rights, competitiveness in party system capable of aggregating diverse interests should
be considered as main indicators for evaluating leader’s performance. Developments
in education (literacy rate) and economic freedom as criteria for industrialization are
also indicators for a nation’s social and economic achievements. Homogeneity index
and population density are more related to govemability that a leader can handle
problems of his own country.

While all variables in personal background attributes may possibly be used in
descriptive analyses in order to compare the leaders across the different aspects, some
selected variables are included in a statistical model based on certain theoretical
perspectives and some variables are deliberately combined for certain roles. The

following diagram (Figure 1) shows the model in a very clear manner.

51

Figure 1 Political Success

 

Duration of Ofﬁce

 

 

 

I
l l l

 

 

 

 

 

 

 

 

 

Demographic Attributes Career Attributes Institutional Attributes
Age ' Political Experience Age of Country
Place of Birth Long Term Foreign Population Density
Experience .
Education (Degree) Homogeneity Index

 

 

 

 

P Identiﬁcation
arty Literacy Rate

 

 

En Mode
try Party System

 

 

 

 

Previous Occupation .
Economic Freedom

 

 

Freedom Status

 

Type of Head of
Government

 

Region

 

 

 

CHAPTER III

Theoretical Propositions

Legitimacy is an idea of central importance in political thought and science
although the concept has proven difﬁcult to use in comparative politics. If a polity lasts
for a long time without disruptive changes, it is considered durable. Furthermore, a
polity’s legitimacy is likely to be enhanced by its longevity. Among the changes, a
change in the formal power of a top executive, or the government as a whole is
considered major. Legitimacy is gained or lost through the performance of incumbent
political leaders. Unless a top executive who fails conspicuously in action as a leader
manages to conceal the failing the political community - as some tyrannical leaders
have done, at least until their death or overthrow - that leader’s authority is likely to
suffer as a consequence of the perceived failure (Tucker, 1981, p.42).

To achieve any lasting improvement in the polity, whether it be economic
development, a sense of nationhood, or an equitable society, a leader must obtain
popular support for his goals, and to do this requires strong leadership. Regime
legitimacy, in effect, goes beyond the performance of a particular incumbent leader and
his or her popularity. The legitimate authority of the government and the support for
the incumbent government is narrowly circumscribed as to its functions and largely
stripped of the personal features of ascriptive authority. The sources of legitimacy
serving as the basis for authority can derive from the forces of tradition, from
charismatic, or from an institutional rational-legal system of government.

People tend to understand regime legitimacy questions largely as conﬁdence in

52

53
particular ofﬁce-holders and what they do (Lane, 1979, pp.63-64). The importance of

legitimacy is considered a precondition for political stability. To a large extent,
legitimacy depends on a social and economic system. In general, personal popularity
of a president or a prime minister over a relatively short period (e.g. 2 yrs) will
contribute fewer positive ﬂows to the stock of regime legitimacy than an economic
crisis over along period (8 - 10 yrs) under different governments will generate negative
ﬂows.

The object is to explain why some political leaders stay in ofﬁce for a short
period and others do not, or why some individuals demonstrate more successful
leadership than others. In every society there are deﬁnable paths to political success.
The requirements for personal success as political performance in office might vary in
a corresponding fashion. ’Years in ofﬁce’ result from popularity and performance
which are multicausal. The reasons can be explained by several components. In the
aspect of social and economic environment of the society, conditions of wealth,
equality, and shared cultural values, etc. inﬂuence the sources of variation in political
performance, that is, duration of ofﬁce. Moreover, the personal qualities that the
leaders are expected to have are pretty different to the people in each country. The
motivation of being or keeping power longer is one of the reasons as well.

In looking at what causes leaders to succeed or fail, one crucial point is how
they use the power that accrues to them in the role of leader. Some use of power is
clearly necessary for a leader’s own survival, as well as to produce changes, as we
witness the fate of leaders who failed to use power at critical times. There also is

personal power that is not necessarily linked to ofﬁce in intelligence, knowledge, and

54

experiences in a speciﬁc area. Some possess a deliberate manner of expression as well
as a way of holding one’s self in public. Rarely be that as it may be, some have
physical attributes such as tall, big, and handsome. While physical factors of stature
are easy to trivialize, they are after real and signiﬁcant determinants of success because
there are fairly clear cultural expectations of physical ideals and what we want our
leaders to represent in relation to them.

The main research is concerned with the relationship between the variances in
successful and unsuccessful leadership characteristics and the variances in what it takes
to be a long-term political leader. All leaders who keep power do not share the same
experience partly because there are all different degrees of power and success, partly
because of the differences between individuals in character and backgrounds, and partly
because of the differences of situation. In an effort to ﬁnd common denominators or
common determinants of the dependent variable, considerable attention should be
devoted to the family backgrounds and early histories, career backgrounds, and
situations.

The major reason for focusing on the elements above is the implicit assumption
that if we know who they are, and were, and how they became political leaders, then
we evaluate what they do. A sequence of innate and inchoate forces within the human
organism epignetically emerge during the development of individuals in their proximal
contacts in families and other naturant groups and in their distal contacts with
economic, social, and political habits, customs, institutions, and values as these develop
pari passu with and within individuals.

The social characteristics of the leaders vary from one political system to

55

another. There are many examples to show how the conditions and requisites for entry
into political leadership vary from country to country and in different periods. It is
because each society has its own social values that political leaders should have and its
principle of legitimacy, which guides the actions of the political leaders and expresses
a body of values and ideas accepted by the people. There are relationships between
personal backgrounds of political leaders and performance of political leaders. Social
origins are important, justifying their focus by arguing that social background is a
useful indicator of political success. Leaders need to possess a large array of personal
attributes if they are to be successful. The career backgrounds of the leaders play a
major role in the determination of policy goals and general orientation of political, life
in a country. Background data on the social and career characteristics of the leaders
provide no infallible guide to either attitudinal patterns or policy preferences.

Leaving aside for the moment such personal attributes as the individual’s ability,
his educational and technical background, what environmental factors inﬂuence leader’s
survival? It seems relevant to consider some important situational factors which
determine whether a person is chosen to be a head of government. While the behavior
of men or of social systems may be described, and perhaps even predicted, without
resource to personality and skills variables, no satisfactory explanation is possible until
we understand the psychological link which joins man to his sociopolitical environment,
and through which each impinges on the other. As a result, for the lack of consistent
or trustful information on personality and skills, the propositions on those elements are
excluded from a theoretical model. Speciﬁc propositions on the basis of social and

career backgrounds as well as situational attributes will be discussed in this chapter.

56
A. Do Social Backgrounds of the Leaders Really Matter for Their Political Survival?

The study of social backgrounds can furnish clues for a study of political
performance, but the ﬁrst cannot substitute for the second (Lasswell and Lerner, 1965).
The analysis of social background data still performs important functions for
socialization and recruitment. A person of a certain age, educational attainment and
social status is more likely to be successful in his office. He may be liked, trusted, and
respected within the society in which he exists. He should have abilities to attract
dedicated followers as well as abilities to command loyalty and dedicated service from
a comparatively small group of intimates.

The demographic variables are based on the concept of naturalness. When
people are born, they get certain social status, gender, age, etc., from their parents.
Parents and schools have emphasized the importance of achieving social prominence
or social dominance. They contribute socialization process to the young children to
learn what an individual is prepared, with varying degrees of success, to meet
requirements laid down by other members of society for his behavior in a variety of

situations.
Proposition 1: Long-lasting leaders are younger on reaching office.

Generally, age forms the basic criteria which determine eligibility for speciﬁc
roles. Without laws to make precise the obligations of office and ﬁx the tenure of
heads of government, it is universally acknowledged that leaders will hold power as
long as their prestige lasts and with little regard to senility. According to gerontocracy

thesis, respect for the aged has been at work in most societies although it has been more

57

common in East rather than in West. In some societies, leadership capacity often makes
its appearance relatively late in life, in the form of a strong personality or a strong ego.
The leader is either an elder or is assimilated to elders even if he is young. However,
the seniority system has its drawback especially in the political arena.

As a result, as far both life expectancy and seniority system are working
together, if the leaders become top executives at middle age, they stay longer than those
at young age. Political leadership correlates positively with middle age (Blondel, 1980).
It would seem to be ’natural’ for leaders to be relatively old With a certain tendency for
the age distribution of the population. However, considering human’s life expectancy,
long leaders are younger on reaching ofﬁce. In this study with continuous age variable,
the proposition should be that the leaders who reached their ofﬁces at younger age
everywhere are more successful.

Proposition 2: . Leaders who were born and raised in urban areas are more likely
to stay in ofﬁce longer.

Political success correlates with the exposure to urban life. This variable
indicates the social environment that the political leader has been raised in. Urban-bom
and raised in urban area inﬂuence the leader get more exposed to politics and media.
The ﬁrst role perceived by children tends to be a top political executive - the president,
prime minister, or general secretary, etc. In early childhood, they have tendency to
identify a top political executive as a parent ﬁgure. So they become more successful

than those from rural or rural-lived.

Proposition 3: Highly educated leader tends to have a longer tenure in ofﬁce.

58

Educated persons are more aware of the impact of government on their lives and
pay more attention to politics. It is because people who have a strong sense of personal
effectiveness also tend to feel politically effective. Personal and political efﬁcacy are
strongly related to education (Campbell, et al., 1960). Intellectuals play a vital role in
formulating, disseminating, or sustaining idea that inﬂuences both the theory and
practice of politics in most societies.

Tucker (1977, p.384) argues that trained and sophisticated insight, the capacity
to judge situations _ accurately that comes from experience and intellect, are
characteristics of an effective leader and are particularly vital requisites for political
leadership in the extremely complex world of today. Formal education is the single
most important entrance requirement into the higher circles. Political leaders are among
the most educated of all occupational groups. The more highly educated people tend
to develop the characteristics of trust, satisfaction, and competence that go with
democracy. It is a common place to note that educational levels are the highest in

socially favored categories.

B. Do Career Backgrounds of the Leaders Have Inﬂuence on Their Political Survival?

Most leaders of the contemporary world have had a ’nonnal’ career. The
majority have been ministers or congressmen before becoming the head of government.
Also, most leaders have been involved in party activities. Competence of the leaders
_ includes references of each person’s past political experience, ability as a head of

government, and his intelligence. Considering the fact that the leaders must be

59

absolutely dedicated to the cause and able to demonstrate that commitment by giving
time and effort to it, risking their lives, undergoing imprisonment, exile, persecution,

and continual hardship, etc. are very inﬂuential to leader’s political survival.
Proposition 4: The greater the involvement in a political collectivity, such as
parliaments or congresses, ministries, and political parties, the
greater the support for the leadership, that is, the political rule of

the game. >

Ambition theory (Schlesinger, 1966) presupposes a rational assessment by the ’
prospective leaders of their opportunities for success. For the most part, political
leaders are far from common men in either their origins or their achievement. Political
experience is the best teacher of the art of politics, and the accumulated experiences of
generations, concentrated into streams of traditions which ﬂow into and through
institutions, such as lineages, professions and schools, colleges and universities, are the
sources of the knowledge which enables political classes to be successful (Shils, 1982,
p.17). If they have experiences in politically related areas such as parties, congresses,
or cabinets, it is easier to ascend to a head of government. To pursue a political career
an aspirant must conform to a recruitment process which rather rigidly prescribes his
role as a candidate for parliament and a member of the ministry (Edinger, 1964, p.663).

Highly experienced public ofﬁcials are more likely to be selected as potential chief

executives.
Proposition 5: The successful leaders are likely to be found in law, politics, civil

service or military, that is, political sectors that would be regarded
as upper-status occupations.

60

The correlation between social status and political power, though consistently
positive, is not a ground for assuming that the former is a sufﬁcient or necessary
condition for the latter (Prewitt, 1970). Also, there is a tradition of respect for the
individual, regardless of his rank and wealth in Western political thought. Nevertheless,
higher social status is necessary in order to have higher political opportunities. Political
job-chances of those with high social status are considerably better than those with
average or low prestige.

The primary occupations of the leaders prior to seizing power or assuming
highest ofﬁce are rather diverse. Also, children can be inﬂuenced by their father’s
occupation in developing their political views. The successful leaders tend to have
higher status fathers from the upper and middle class holding such positions as landed
gentry, big businessmen, government ofﬁcials, lawyers, military ofﬁcers, and politicians.
However, the leaders had more politically related professions than their fathers did.
While fathers’ occupations provided better chances (high social status or wealth) to
enter the realm of politics, the leaders themselves established their careers and used
them as stepping-stones to the highest political jobs.

Proposition 6: Leaders with foreign experiences, i.e., foreign education and work
in foreign countries, are more likely to stay in ofﬁce for a long
period.

Foreign experience contributes the leaders ability to develop cosmopolitanism.
Abroad, they see different level of development and understand of foreign affairs. Most
leaders will tend to heighten their cosmopolitanism by greater foreign travel, foreign

study, exposure to various societies and cultures, and maintenance of foreign contacts.

61
They are guided by the idea of adopting and adapting modernized technology through

foreign education or foreign experiences. They contribute to making society stable, and,
of course, successfully accomplish more. General knowledge is associated with foreign
language ability and with understanding of foreign affairs. However, this tendency
depends on the level of development of country or reign of country.
Proposition 7: If a leader comes to his ofﬁce in a peaceful way, he is more likely
to stay in ofﬁce longer.

In general, some political leaders come to ofﬁce as a head of government
through peaceful processes, whether appointive or elective, while others in violent ways.
Some scholars (Rajai and Phillips, 1988) identify these two groups of individuals as
follows: The former are ’loyalists’ and the latter are ’revolutionaries’. Since, in this
study, just the leaders in single leadership would be discussed, entry mode as an
appointment mechanism can be a variable to inﬂuence political leadership in republics,
presidential or military presidential, and prime ministerial systems. Even though party
secretary or general secretary in Communist system is not real single leadership, they
are normally stronger than the prime minister.

Political leaders, having been chosen and exercising their ﬁmctions in
accordance with these rational-legal procedures, can enjoy a high degree of legitimacy.
In addition, they need not worry about being humiliated, executed, put in jail, or forced
into exile after an overnight coup. Removal from ofﬁce through constitutional due
process or assassination are the major and constant risks to their political survival. If

legitimacy is deﬁned in terms of established rules, a government which has come to

62

power by civil war, revolution or coup d’etat can not immediately claim legitimacy,
although it may claim authority by virtue of its ideology, goals and popular support, by
providing its effectiveness in meeting the country’s needs and by demonstrating the
qualities of the new leaders (Carter, 1979, p.51).
Proposition 8: Leaders of the right, such as the conservatives, tend to stay longer
than those of the left, such as the liberals.

While the political left has higher levels of tolerance for democratic liberties
than those on the right, the political right is less supportive of democratic norms. It is
important to pay attention to the beliefs and values of political leaders, particularly as
they relate to sweeping ideological and intellectual movements. Political leaders always
possess some sets of philosophical and political convictions that are his own. In order
to legitimize any type of regime, the leaders of the regime must devise a political
formula and political ideology that gains acceptance among the people at large.

Ideology plays a signiﬁcant part in the life of most parties, both symbolic and
practical. They try to shape political institutions in ways consonant with his own
political philosophy, even though the goals of the leaders are not necessarily identical
to those of the followers and intermediate leaders. They generally identify themselves
with the ideologies7 of the right (conservative), the center, or the left (liberal) in terms
of their stances toward the size of government, the equalization of social conditions, and
the social and economic issues in a country.

Ideologies of the left have tended to favor a larger role for government, greater

 

7 Right means traditional conservative and populist (even if it is a kind of left).

63

equalization of income, govemment-supported welfare programs, and a minimization
of unemployment. Meanwhile, the right has tended to favor a more restrained role for
government, maintenance of existing income distribution, private welfare, and a
minimization of inﬂation, even at a cost to employment. Usually, the more
conservative the ideology of a political leader, the longer he stays in ofﬁce. It is
because social and economic resources based on conservative ideology induce
politically stable situations. The promise of radical change in the social, economic, and

political fabric, is threatening to political stability.

C. How Do the Institutional Attributes Inﬂuence Political Survival?

Geographical environments, economic systems, social structures and cultural
differences provide some assistance to explain political success on a macro level.
Geography and demography would be considered as physical attributes of countries.
Thus, size of country“, population size, and age of nations might be the main indicators
for both attributes and these variables imply govemability or viability, legitimacy of the
countries, or vulnerability of the nations. The duration of the country, the unity and the
characteristics of the political culture are the indicators for national identity. It is
difﬁcult to handle and almost impossible to quantify.

Proposition 9: Leaders in densely populated countries are more likely to stay in
ofﬁce for a short period.

 

8 According to de Blij (1967, p.41), states exceeding one million square miles are generally
described as very large, while those under 10,000 square miles are referred to as very small. The
category of the size of nation is as follows: Very small ( - 10,000), small (10,000 - 60,000), medium
(60,000 - 140,000), and large (140,000 - ).

64

There may be considerable variation in the potential contribution of geographical
elements in achieving the basic objectives that usually underlie state interventions: (1)
the provision of security and order, related to the defense of a territory; (2) the
stimulation of economic growth and well-being; and, (3) the guarantee of social justice,
etc. Geographical space can fulﬁll the realm of government actions, demonstrating the
rationale of functions on the grounds of either efﬁciency or equity criteria. In addition,
the application of policies requires some spatial considerations for the same purpose to
get efﬁciency or equity.

The problems of size as related to effective national control and politico-
territorial organization are nothing new in the history of state evolution. As seen in the
literature review, Montesquieu already mentioned the regime types based on the size
of country. According to the previous cross-national studies, population size has an
inﬂuence on government performance or government stability since it affects overall
rates of political change. As a result, the more densely populated countries are more
unstable because high level of population density becomes a main source of change and
social dislocation.

Consider the redistributional element of public programs, where the spatial
distribution of beneﬁts to individuals and household ought to be in accOrd with
population - which is in fact the case for education, health, housing, and income-
maintenance beneﬁts. Governments of small countries have to deal with qualitatively
less heterogeneous societies - similar attitudes, values, and beliefs. In a general sense,
a chief executive in a country with small population is more likely to be successﬁrl in

ofﬁce.

65

Actually, country sizes run the gamut from the minute Vatican - less than one-
sixth of a square mile - to that of the giant former Soviet Union, over 8.5 million square
miles prior to the collapse of the Eastern Block. With respect to country size, it is
easier to keep order in smaller nations (Lijphart, 1977; de Blij, 1967). In a small nation
it is easier for authorities to perceive when discontent is smoldering and to act to
prevent its outbreak, either through suppression or by alleviating its causes. Meanwhile,
in large nations, the information processes work less well, and control is more difﬁcult;
hence, outbursts are more likely. Small countries offer better requisites for the creation
of the spirit of cooperativeness and an accommodation, which is crucial for the
maintenance of democracy (Lijphart, 1977, p.65). In such countries, it is probable that

the political leaders will be more united through personal acquaintance and interaction.
Proposition 10: In older countries, leaders tend to have longer duration in ofﬁce.

As one of the indicators for legitimacy, age of nationhood is very important.
An older country is more likely to have achieved a high level of national identity than
a newer country. An older country is stronger, having usually lived through a number
of crises. The political leaders in the older countries are likely to be more successful
since those countries tend to have institutional stability and political tranquillity. It'is
because opportunities for succession are more numerous in older countries.
Irregularities of succession are rare among very new countries.

This proposition is pretty much related to colonial experience of the country.
Colonial experience in each country’s history reﬂects the establishment of her own

background. Even though one country is located in a certain region and has been

66
sharing same cultural inﬂuence in that region, if she has been affected by foreign

power, her own culture should have been changed in some degree. The new states
were handed down the political and administrative structure of the imperial nations.

For example, the British built up a fairly well-functioning administrative system
in the colonies, imposing a kind of indirect rule. France, meanwhile, applied a
markedly centralized administrative system, which was sustained in all essentials by
ofﬁcials from the home country. Thus, the population of the French colonies became
less involved in the exercise of authority. As a result, the states which the British left
behind would, for these reasons, be better equipped for democratic government than
those which had belonged to France. The impact of colonialism on the traditional
societies of Asia and Africa is signiﬁcant.

Low modernization is associated with low participation, low competitiveness,
and fewer political parties. Lower levels of modernization are related to a party system
which is relatively less complex and characterized by low competitiveness where a
single party has been established as the dominant one (Huntington, 1968). As
indicators of structural attributes, we may use party system, economic. system, and
cabinet size may be formally used.

The issue of political success is related to whether the political leaders have
legitimacy and efﬁciency. An efﬁcient institution of a chief executive will become
more legitimate as time passes, while one which is inefﬁcient will lose credibility. To
keep a position as a prime minister or a president, an individual must have a strong
desire for office and a willingness to do whatever the system requires to reach the top.

If legitimacy entails the involvement of many groups in the selection of the chief

67

executive as well as in the decision-making process, the result may be a reduction in
efficiency through instability or indecisiveness.
Proposition 11: The more fragmented the party system, the more unstable we expect
the government to be; and the shorter the leader stays in ofﬁce. As
a result, government stability is negatively correlated with the
number of parties in the government.

Another line of inquiry focused on the impact of the party system in maintaining
political stability (Pridham, 1990). The stability of political leadership depends on party
activities in the electoral and legislative arenas. Political parties play a leading role
everywhere in legitimizing and institutionalizing the paths of political careers, not only
as an instrument for winning and administrating power but also as an instrument best
adapted to political struggle. We know that some parties adapt very easily to open
competition and the peaceful transfer of political power while others are exclusive and
seem unable to react to power alternatives, with the exception of violence. The more
diffuse are the party roles, and the more variegated the career paths, the more diverse
the people selected thereby. Parties are two-way means of communication of inﬂuence
between rulers and population.

Party system is related to the degree of national integration and dimension of the
process of national development and modernization. Economic development is more
likely to be realized, given one dominant political party, or a pair of parties, each
strongly committed to economic development, as opposed to a multiplicity of parties
(Lerner, 1957). The number of parties affects democratic stability (Lipset, 1960;

Blondel, 1968; Tylor and Herman, 1971; Sanders and Herman, 1977; Dodd, 1976;

68
F isichella, 1984). If party systems are highly fractionalized, government tends to form

in unstable and conﬂictory political systems.

Competitive party systems ensure relatively high levels of turnover in
government (Mishler and Hildreth, 1984). A multiparty system tends to depress
people’s support for a leader. Meanwhile, a single party system sees itself as
synonymous with the entire people and offers less political opportunity and higher
political risk than parties in a competitive system. Consequently, leaders in a single
party system or a single-party dominated system are more likely to stay longer in office
than their counterparts in competitive party systems. A single party system usually
leads to a totalitarian system in which the government nevertheless strives to achieve
mass political support on the basis of a powerful, visionary ideology.

However, a stable multi-party presidential democracy is more difﬁcult to achieve
than a two-party presidential democracy or a multi-party parliamentary system (Lijphart,
1992, p.116). Countries with at least one effective governing party are more likely to
deal successfully with the problem of succession than are countries where parties are
not present. While governing juntas of bureaucrats and military men are rarely prepared
for the politics of peacefully electing their own leader or of having leadership
peacefully pass from one person to another, at least parties are experiences in the art
of internal elections.

The strength of environment promotes leader’s success. For example,
prosperous economic periods give heads of government more chances to obtain their
success. Economic development fosters democratic political stability; the state’s

capacity to promote economic effectiveness is itself a key ingredient in Lipset’s

69

formulation of stability. The economy’s infrastructure such as wealth, industrialization,
education (literacy), and urbanization, leads to the emergence of viable political
performance (Lipset, 1981, pp.27-86).

Long leadership tends to exist in countries where the extent of social and
econorrric change is low. If the freedom of social, political, and economic areas is
guaranteed, the changes will be less. Thus, the society tends to very stable and the
political leader will stay in ofﬁce longer. Economic freedom, freedom index as a
combination of political right index and civil right index, homogeneity index,
urbanization, and literacy are used as indicators of economic and sociopolitical
development for the hypotheses testing.

Proposition 12: The lower the extent of social and economic change, the longer the
leadership will tend to prevail.

The level of economic freedom depends on what economic system the society
is in. While economic systems vary considerably among cultures, they can generally
be analyzed by looking at who manages the system and how, and upon what basis the
goals are distributed. Democracy is associated with a certain form of economic
organization, namely capitalism. Capitalism promoted a dispersion of power in society
which creates a favorable basis for political democracy.

Meanwhile, under socialism, a strongly concentrated, organizationally coherent
economic and political power breeds passivity and fear among both the citizens and
potential opposition parties. No one dares to seriously resist the political leaders who

control his livelihood (Schumpeter,l975). Especially, in the turbulent politics of

70

developing nations, leaders can never take tenure for granted. Political survival must
be actively pursued by manipulating public policy to construct supporting coalition.
Proposition 13: More divisions or cleavages of race, language, and religion in
society increase instability. Therefore, the leader who is in the
more homogeneous society is more likely to stay in ofﬁce longer
since the society is more likely to be stable.

Ethnic, linguistic, and religious differences are inescapable facets of politics
within as well as between countries; internal conﬂict and instability ﬂow from tensions
and processes of mobilization. In deeply divided societies (Lorwin, 197; Nordlinger,
1972), increased political consciousness, as the result of social mobilization, could be
expected to strain or destroy the cohesion of states (Deutsch, 1961). The homogeneity
index implies a measurement for social and political cleavages. As cleavages increase,
so does political instability. A lack of homogeneity in the population contributes to
increased political violence, i.e., political instability. If all population had the same
political choices, the task of political job would be much easier. Where social
cleavages are deep, the legitimacy of the social arrangements is low, and leadership is
less likely to emerge unchallenged.

Political cleavages result from a set of attitudes that divides the nation’s citizens
into major political groups based on religion, language, race and tribe. Those cleavages
represent a sense of distance between factions and causes fear, distrust, and hostility
between them. It becomes the source of major conﬂict and leads to low integration.
To that extent, loyalty to country or a head of government itself is dispersed. Political

tensions and conﬂicts arising from social cleavages can seriously affect national

71

political instability and the capacity of governments to control their territories. As a
result, the greater the homogeneity in a society, the more executive stability given the
problem of issue bargainability. In a democratic society - religio-ethno-linguistically
homogeneous - or alternately which provides a great deal of economic mobility - large
degrees of economic inequality will be tolerated.

Proposition 14: As one of the social mobilization indicators, a higher literacy rate
results in a higher level of political stability, and, therefore, the
leaders in that society stay in ofﬁce for a long period.

Countries with modernized social structures and developed, industrial economies
are expected to have a higher level of stable and effective government, for the
indicators of modernization, urbanization and literacy rate are important (Lerner, 1957;
Lipset, 1960; Deutsch, 1961). An urban and industrial society makes democracy
possible by providing a fair distribution of wealth, a public with a sense of political
efﬁcacy and the time and inclination to participate in politics. Literacy and urbanization
pertain to a development away from a life of local isolation, traditionalism and political
apathy, and into a different life of broader and deeper involvement in the vast
complexities of modern life, including potential and actual involvement in mass politics.
However, urbanization is not included in the test since it may be highly cerrelated with
populated density.

The literacy rate is indicative of the level of aspirations and expectations in
society. It broadens outlooks, increases tolerance, reduces extremism, enhances belief
in democratic values and practices, develops a capacity to make rational electoral

choices, and reﬂects greater skills and specialization (Lipset, 1960). Greater literacy

72

promotes openness and a deeper insight into political issues and contributes to increased
participation. The literacy rate, by specifying the proportion of population which can
read in one language, gives a fair picture of national participation in the whole cultural
sector. According to Lerner (1957), literacy around the world correlated at .84 with
per-capita income and at .87 with industrialization.
Proposition 15: Leaders in the countries with high levels of freedom status are more
likely to have longer durations in ofﬁce.

As far as a deﬁnition of democracy can be incorporated into a concept of
stability or institutionalization, which refers to the degree to which the political system
may be expected to remain in existence, political democracy is associated with freedom
of the individual. The degree of repressiveness imposed by a government indicates not
only the extent to which a system of governance will tolerate its political challengers,
but the plan of a coercive compliance structure. Political freedom is a combined index
that political right and civil right indices that are measured by a govemment’s
regulations which deprive basic human and political rights. If a society does not assure
a degree of equality and participation, pressures that may undermine stability can be
generated. The cost factor of political participation is directly linked to the extent of
a govemment’s coercive activities, that is, the probability of getting caught as well as
the severity of punishment.

In a nondemocratic nation, that is to say, a nation with low political rights and
civil rights, even though it is more likely to be unstable, the leaders can hold power

long since the government sanction is too extreme and the cost of participation is too

73

high. However, sometimes there have been exceptions: for example, Pol Pot in
Cambodia, Amin in Uganda, Marcias in Equatorial Guinea, and Bokassa in Central
Africa, all ignored social and economic policy-making because they were able to avoid
pressure for social and economic development only if the aspiration of their citizens
were extremely low (Blondel, 1987, p.71).

Nonetheless, even in a society with high political and civil right indices, where
political institutions are relatively weak, leaders also can not stay longer, because the
cost of participation is low and leadership is more likely to be vulnerable. The leaders
in the society have less political power over the nonleaders. It increases the likelihood
of political instability by exacerbating political divisions and conﬂicts and by
encouraging criticism of the incumbent regime.

Proposition 16: A presidential system based on elections has a longer duration of
ofﬁce than prime ministers through selection systems. In general,
presidential systems are more stable - as measured by the tenure of
the chief executive - than prime minister systems. The greater and
more autonomous presidential resources make such stability likely.

The basic constitutional rules by which political leaders are chosen and
authoritative decisions are made should be considered. They can be operationalized by
systems of executive leadership or institution of the executive. It explains the
executive-legislative relations how independent a chief executive is from congress.
Even though the position of prime ministers or general secretaries as collegial
leadership is not as strong as the position of the president, they are still the most
prominent ﬁgure in certain ways. Selection results in nomination to ofﬁce, while

election results in the holding of ofﬁce. Prime ministers are less personality-oriented

74

and less focused on public image. In collegial leadership, the members of party or the
members of politburo elect a top executive.

Proposition 17: Cultural patterns based on regional differences, once established,
possess considerable autonomy and can inﬂuence subsequent
political and economic events, i.e., democracy (regime type) and
economic development. Political culture plays an important role in
the development and maintenance (or failure) of democracy.

Based on regional comparisons, cultural difference should be considered as a
control variable. Countries which have a common culture can be categorized, and these
categories used in comparisons will then be likely to carry the same or similar meanings
in each political system, such as Latin America, Europe, Asia, Oceania, the Arab
Middle East, and Africa. Each political culture, as well as each region of the world
differs somewhat in the typical backgrounds of its rulers. Political culture refers to
widespread ideas, habits, norms, symbols, and practices that are politically relevant to
a given system, especially as they relate to the legitimate use of power. Since the
traditional culture of the period stressed obedience, respect, and deference to authority,
there was a tacit general acceptance of the principle of government by men rather than
laws.

Alongside legitimation based on public consent, there are vestigial expectations
about political leadership which are derived from the mixed Western heritage of
monarchy, aristocracy, and the like. For example, while the path to democracy in
EurOpe and the Anglo-American nations that were protestant resulted in higher levels
of economic development and stabilization of democratic government, Asian countries

followed Confucianism as an alternative road to democratization (Diamond, 1993).

75

However, one would hardly expect to ﬁnd either democratic cultures or stable
representative, liberal, constitutional regimes in Latin America or Africa. Without the
Protestant tradition, there are likely to be unstable or perhaps ephemeral. The
authoritarian, hierarchical, monistic cultural heritage. of Latin America has greatly
diminished the prospects for liberal democracy (Inglehart, 1990).

If there were cultural differences in the perception of political leaders and
leadership, then one might reasonably anticipate that the cross-cultural validity of the
ﬁndings associated with research on implicit leadership theory would be affected.
Different cultural values affect the acceptability of ideas and practices associated with
political leadership across the nations. Dispositions toward authority are different in
each political culture.

Western democratic political culture denies faith in powerful leaders, that is,
authoritarian personalities, while common tendencies in Asia are more focused on
loyalty to the collectivity over individual freedom and needs as well as on respect for
paternalistic authority relations (Pye, 1984, pp.184-185). Even after the independence
from a colonial control, cultural renewal is ﬁrst and foremost the job of top political
leadership. Cultural values should be exempliﬁed by top leaders to provide supreme

examples for others to follow.

CHAPTER IV

Operationalization and Methodology

In the theoretical parts in the previous chapter, 17 propositions relating to a
leader’s political survival were presented. In this chapter, ﬁrst of all, data collection
and sample selection will be discussed. Second, all independent variables and a
dependent variable for testing the 17 hypotheses will be discussed and operationalized.
Finally, with those operationalized variables, the most appropriate method for the

research and its interpretation will be introduced and discussed.

A. Data Collection and Operationalization

Nation states arise from the time that they achieved independence as sovereign
political units. The victory of the democracies in 1945 seemed to open a new era for
political progress, with the United Nations as the symbol of the optimism. Between
1945 and 1991, almost 98 former dependencies joined the ranks of independent actors
on the world stage. Nonetheless, of the approximately 175 independent countries that
exist in the world today, more than 130 governments can be considered nondemocratic.

The popular political system, that is, presidential or parliamentary, can not be
applied to all independent countries. The large majority of third world countries
function under a system which does not fall within the province of these two systems
of government, and, deﬁnitely, communist countries operate under quite a different
model. In modern constitutional regimes, a chief executive is chosen through each

constitutive system. Some become presidents by referendum in general elections, some

76

77

by national assemblies, or some by coups or violence. Along similar lines, the
communist leaders, i.e., the General Secretaries, are elected by the party in a one-party
system, whereas the prime ministers also would be the heads of the government by the

elected assembly or by a chairman of the party in the parliamentary system.

Data Collection

The sample in this research is composed of the political leaders who actually had
been heads of government from 1945 to the end of 1991. 997 leaders from 147
countries including those who were still in power in 1991 when the data were collected
are included. Among all the heads, the role of a head of government is most important
in decision-making. As mentioned earlier, since ruling monarchs have gradually shrunk
since World War 11, those are excluded from the samples. Also, all collective regimes
are eliminated.9 However, even stop-gap, interim leaders, who are the truly temporary
appointment of men chosen because they hold a neutral position in the political
spectrum, are included in the sample. They usually would continue to be in these
ofﬁces since they used to be vice presidents or deputy prime ministers before getting
into interim positions.

Considering both left and right time limits, some manipulations for data selection
are needed. If a leader had entered his ofﬁce before 1945 and got out of his ofﬁce
before 1992, he or she is included in the sample. However, although a leader became
a head of government during that period and was still in power, it is not certain whether

he can not be included in a target sample for the analyses since it is not possible to

 

9 See p.43 in CHAPTER II.

78

Figure 2 Distribution of Duration (Percentage)

 

 

 

 

 

(x)
25
All Leaders
C997)
20
Leaders out of Office
(882)
15
Leaders In Office
(11S)
10
5
0 14411le

 

 

 

 

0 2 4 6 B 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 4O 42 44

Duration (Years)

Figure 3 Cases Excluded from the Sample

 

 

   

 

Excluded from the sample

Total cases are 13.

 

 

.—

lnciuded In the anaIySlS

 

 

 

 

 

 

 

 

 

 

 

 

m-

 

1945 1989 1990 1991 1992

79

know his exit mode and correct duration of ofﬁce. The plot of the frequency
distribution of durations seems to be an exponential speciﬁcation showing that it reveals
a large concentration of shorter durations with the numbers of cases diminishing as
durations become more lengthy (see Figure 2); high percentages of the distributions of
0-2 years exist in all leaders as well as the leaders out of ofﬁce. In order to reduce
missing data that represents the duration of leaders (115 leaders) who were still in
ofﬁce, 13 leaders who were in the duration of 0-2 years are deleted from the whole
population. In other words, the leaders excluded from the sample are those who got
into ofﬁce since 1989 but just had been in ofﬁce for less than 3 years. The leaders who
were still in ofﬁce, but had been in power longer than 2 years will be included in the
statistical analyses (102 cases).

Although longevity of a top executive has been theoretically explained by
governmental stability or durability, it is, in a direct sense, a matter of personal survival,
not governmental durability. To measure the length of his service is to measure the
length of time that any one person holds ofﬁce at a single stretch. Therefore,
discontinuous terms of each individual leader are counted separately. Similarly, each
regime of a country can be taken as a separate historical instance, i.e., as an individual
case. When a national election is held, if an incumbent is reelected, the government is
assumed to be the same government as the previous one. However, if a same leader
became a top executive not in a consecutive terms, he is coded as a different case.

There is nothing inherent in nonquantitative methods of political analysis that
would prevent their application to the study of political leadership. The discovery or

neglect of political leadership as a subject for political inquiry could not be attributed

80

primarily to the presence or absence of a commitment to quantitative scientiﬁc methods.
New methods and facilities for collecting quantiﬁable data have provided scholars with
an unexpected amount of information in the area of leadership research in part as well
as in the modern societies. However, its selection of data has to be based on a
previously established general theory. It should keep an eye on not only the problems
of relevancy, but also the problems of validity. In fact, the problems of
conceptualization and Operationalization are not easy.

In most research on political leadership, biographies have been regarded as the
main data sources. A political biography is a case history which should move us
toward comparative generalizations about the success or failure of similar individuals
in different situations and different individuals in similar situations (Edinger, 1964,
p.671). However, the use of the biography is vulnerable to scholarly criticism. For
example, according to Sidney Lee (1928):

....True biography is no handmaid of eternal instruction. Its purpose is

not that of history. It does not exist to serve biological or anthropological

science. Any assistance that biography renders these three great interests

- ethical, historical, and scientiﬁc - should be accidental; such aid is

neither essential nor obligatory.

Subjectivity of biographical use results from several aspects. First, unsuccessful leaders
tend to be slighted, because material on their lives is frequently far less readily available
than it is for their more successful rivals. Second, the operational criteria seem to be
determined as much by the author’s personal inclination as by the cultural climate in
which he writes.

Most political biographies have collected information from several sources. The

81

directories with short articles on each man are more inclusive in their coverage but
contain only a minimum of information. The best sourCe of information about the lives
of political leaders is the popular biography. The daily, weekly, and monthly press both
in the United States and abroad contains a tremendous amount of biographical
information.

Some popular biography is indexed and readily available to the researchers.
Also, although not all materials are indexed, some biographical information is most
available in newspaper obituary columns. Since the evaluation of their performance is
different, their backgrounds might not be recorded in an objective manner. Even
though the use of a number of different sources yields conﬂicting facts and conclusions,
those are, in fact, minor problems.

Rather, political biographies have tended to fall into certain categories which are
marked by how much signiﬁcance is attributed to the subject’s personality or to his
socio-political environment. The creative or literary biography is more deliberate in the
selection and organization of data. The data on personal background and skills and
personality of the political leaders can be collected from several kinds of sourcebooks
(see APPENDIX A).

For example, a Who ’s Who type of source is descriptive material about the
incumbents of elite positions from which the more comparative studies attempt some
sort of conﬁgurative analysis of elite attitudes and behavior. The gathering of social
background data on political leaders should go hand in hand with the examination of
the historical, situational, social, and psychological conditions under which they have

emerged. These endeavors involve close scrutiny of primary and secondary sources

82

with heavy reliance on biographical materials, obituaries, and historical dictionaries, to

the extent that these are available.

ngrationalization

In a strict sense, the duration of political ofﬁce as a head of government takes
place within an institutional framework which is determined in principle. Even though
it is determined by an institutional or structural (constitutional) aspect, in the behavioral
aspect, it is dependent on the effectiveness or performance of political leaders. It
provides a test of the extent to which the arrangements are applied in practice and of
the indication of the strength and authority of both the chief executives and of the
political system. And, it suggests whether certain types of arrangement are generally
more successful than others.

The duration of chief executives in ofﬁce is the most single important
characteristic for evaluating their survival or success; it relates in part to the structures
of the ofﬁce itself, as some forms of government restrict the period during which the
chief executive can stay in ofﬁce. At least, we can compare the characteristics of the
leaders who have succeeded in staying in ofﬁce for one or more decades with those of
’ordinary’ heads of government in terms of their social and career background and their
environment. In measuring a duration of ofﬁce, if the duration is less than 6 months,
it is counted as 0, and whenever the duration includes more than 6 months with some
whole years, it is rounded off to a year. As a result, the unit of measurement is ’year’.

So, it is possible to obtain information on the composition and duration of the

political leaders and, thus, have a clearer picture of the event and type of experiences

83

which are shared by the rulers of different countries. Qualitative data on the
backgrounds of political leaders can provide information about their origins and
experiences. Biographical or historical materials are usually quantiﬁed as nominal,
ordinal, interval, and ratio scales. A nominal scale, however, does not properly involve
a quantitative dimension, since its numbers serve only to identify qualitatively distinct
entities. The social background variables such as age, birth place, and educational level,
as well as the career variables such as political experience, previous occupation, foreign
experience, and party ideology, will be tested for the hypotheses.

Proﬁles of political leaders do not yield any common threads. Institutional
settings are so dissimilar that we are unable to formulate a set of common assumptions
that apply to every situation. There are no precise solutions to a process that seems to
be based on the personality of the individual, the collective culture of a group, and the
situation at hand.

Keeping attention to the question "who should rule?", it is important to think
about what principle of selectivity should be used to choose political leaders. Who
becomes a political leader is not just decided by elections, primaries, and voting but,
rather, is also the result of a continual sifting and sorting of the citizens who enter the
quest for political power. Furthermore, there are many examples to show how the
conditions and requisites for entry into the political class vary from country to country
and in different historical periods: Wealth, birth, and individual merit.

The social background study provides possibilities to salvage the postulate of
class interest by applying it not to individuals but to large numbers. It is also possible

to shift the emphasis from the social or class origins of political leaders to their

84

education, training, and other formative experiences. Comparisons over time and
among countries can greatly help in the formulation of such suggestions and
hypotheses.

The variables in this group are mostly personal backgrounds, that is, social
background variables and career variables. These are very inﬂuential to political
orientation on each area. For examples, family socialization depends on the social
status of their fathers, the occupation of their fathers, and the place of birth. School
socialization depends on educational backgrounds, major at schools and involvement
in activities at school. Other sources for political orientation are traveling abroad,
career attributes and military experiences, etc. Age, gender, and residence often have
as great an inﬂuence on political opportunities as the class structure.

The strength of a leader depends on various characteristics of a purely ’personal’
character. A political leader is almost always possessed of some set of
philosophical and political convictions that he has, for one reason or another, made his
own. As much might be said of almost every political leader who makes any pretense
at ideological sophistication. For the most part in the studies of political leadership
(psychological approaches), traits have been based on the (often expressed) assumption
that the same skills and traits are required to be leader in all situations and
circumstances. Desire, skill, and endurance are most necessary to succeed in a political
career before someone can expect to be a political leader.

Some ad hoc procedures deﬁnitely will be necessary to estimate the values of
the explanatory variables at the times of events. For the institutional or country data,

the time when the leaders became heads of government should be considered carefully.

85

On the other hand, some country variables such as population density, urbanization,
literacy, freedom index, etc., are considered as time varying, others, such as country
size, cabinet size, party system, economic system, homogeneity index, etc., are
relatively constant or the same over time. Time varying variables are usually measured
at the year when the leaders became chief executives. The sources of data collection

and all variable lists for the research are in APPENDIX A.

B. Methodology
Statistical Analysis

When all units are in the data set and all explanatory variables are observed, yet
some values of the dependent variable are unobserved, those unobserved cases are
censored intentionally. To drop the censored observations from the analysis is not only
to discard useful information, but also to create predictable biases. Thus, in an effort
to minimize the missing speciﬁcation errors as much as possible, manipulation for data
selection was introduced earlier in this chapter.

Also, if the dependent variable (Y), duration of office, is considered as a
continuous variable, a linear regression model may be possible. However, for certain
theoretical purposes, the dependent variable is divided into 4 categories, and if it is
limited in some way, the bias that arises from using least squares becomes a serious
problem. Using regression estimates with a qualitative dependent variable can lead to
serious errors in inference (Aldrich and Nelson, 1984, p.9). The least squares criterion
is inapprbpriate or difficult to apply, since it can not easily be applied to variables that

are ordinal or categorical in nature.

86

As seen in the previous theoretical part, a theoretical model was introduced
based on ’duration of ofﬁce’ as a dependent variable and three groups of exogenous
variables, such as social, career backgrounds, and institutional characteristics, that
included up to 27 dummy variables and 5 continuous variables (see Table 1). Thus, the
analyses at ﬁrst, will be preceded by three reduced models in order to test how each
exogenous variable group relates with a dependent variable, and ﬁnally, by a grand
model to examine the interactive effects of social and career backgrounds as well as the
institutional characteristics.

Based on the individual unit of analysis, a theoretical model with ’duration of
ofﬁce’ will be tested by a two-equation model that comprises a selection model
equation including 984 cases with a two-category dependent variable, and a substantive
model which will be applied to only 581 cases which are selected as a ’NON-SHORT’
category of the dependent variable, ’SURVIVAL’, in order to test the strength of political
survival.

The data for this research, as was mentioned above, are not from a random
sample. Instead, the research attempts to examine the entire population of relevant data
for the theoretical perspective, and, therefore, the violation of one of the assumption
underlying a nonlinear probability model (probit or ordered probit) should not be a
concern. Probit parameters are estimated by a method called Maximum Likelihood
Estimation (MLE). MLE is different from OLS in that it does not have unbiasedness,
efﬁciency, and normality with a small sample. However, in the case of large sample,
it exhibits the asymptotic properties of unbiasedness, efﬁciency and normality. The

main goal in using a two-equation model is based on stochastic censoring. The two-

87

equation model is speciﬁed as a probit in the selection equation and an ordered probit

in the substantive equation.

Selection Equation:
As a dependent variable, SURVIVAL (Y,,) is a binary variable. A probit model
is appropriate for a test using a dichotomous dependent variable based on the Bernoulli

distribution. Yli has two values (0,1) and there will be only one threshold (u=0).

Yli = 0!. + 251:in + “i ui~ N“), 521)
YIi = { 0 (SHORT) if 0 5 duration 5 2
1 (NOT-SHORT) if durationz 3 .

Unobserved theoretical dependent variable, Yi', satisﬁes

Yi. = :kaki + ui “i” N(Oa 521)

where the independent variables (Xis) are as followings (See Table 1):

xI = AGE x, = POLTWO xl7 = CONSERV x,, = PARTFREE
x2 = COLLEGE xlo = POLALL x., = OLD x,, = FULLFREE
x, = URBAN xII = FORSTUD x., = PREZ x,., = ECONFREE
x, = LAWPOL XI, = FORWORK x20 = SINGLEP x,, = ASIA

x, = CIVSERV x., = FORBOTH x,. = MULTIP x2, = AFRICA

x, = MILITAR x., = NCONSTI x2, = HOMOGEN x,o = MIDEAST
x, = LIBTECH ‘ x,, = LIBERAL x,, = POPDEN x,l = EURNAME
x, = POLONE x,, = CENTRE x,, = LITERACY x,2 = LATIN .

In this formulation, the function of probability of getting value Yn = 1 (NOT-SHORT) is

the cumulative distribution function (CDF) of the random variable Y,’.

88
P(Yli = 1) = P(Yi. > 0) = P(ui > 'szin)
= P(ui < zﬂkxki) = P(ui < Zr)

= F(Zi)

Zi
1 J 8.8/2 dt

 

= V21t

of!)

To obtain an estimate of the index Z,, the inverse of the cumulative normal distribution
function, F“(P,), can be applied.

The selection equation with a two-stage procedure is to estimate the probit
function determining whether a leader stayed in ofﬁce ’SHORT (Y,,=0)’ or ’NOT-SHORT
(Y,,=l)’. Basically, for this function, we need to assume that we observe Y2,
(STRENGTH of survival) when Y,,> 0. The main goal of this function is to get a value
of A, (hazard rate) for each observation in a substantive equation, which comes from the
sample selection procedure. Using the selection equation, we get an estimated value
of Z,=-BX, for every case and calculate a hazard rate (A,) by ¢(Z,) / l - <D(Z,). Then,
we use Ai as one of the explanatory variables in a substantive equation.

The A, has a similar ﬁmction as a z-score and monotone decreasing function of
the probability of the observations selected. It is the probability that observation 1 has
data for Y2, (STRENGTH). <D(Z,) and ¢(Z,) respectively are the density and distribution
function of the standard normal distribution. The hazard rate has the following two
characteristics (Heckman, 1976, p.479). First, its denominator is the probability that an
observation i has data Yi ; secondly, the lower the probability that an observation has

data on Y, the greater the value of A, for that observation.

89

Substantive Equation:

In a substantive equation, STRENGTH (Y2,) is used as a dependent variable. As
mentioned above, 581 cases in the NOT-SHORT category of SURVIVAL are analyzed with
3 categories: 1 (3-6) with 357 cases, 2 (7-12) with 131 cases, and 3 (13 +) with 93
cases. Even though the same independent variables are included in this model, the
hazard rates obtained from the selection model are used as one of the independent

variables in the substantive model. The model speciﬁcation is as follows:

Y2i = 0t + Zkaki + ui “i” N“), 521)
Yi. = o. + Zﬁjkxjki + “i ui~ NU), 521)

y,,= { 1 if u,.,.< Y,'Sp.,
0 otherwrse

for 1= 1, , n
j = 1, , M
M = The number of categories

1 (3-6) if Y,°<0
Y21= { 2 (7 " 12) if 0 S Yi.< “'2
3 (13+) if LIZSY,’
* ( ) indicates the duration of ofﬁce for each category.

Since ’STRENGTH’, like ’NOT-SHORT’, is divided into 3 categories, the linear
regression model can not be used since the dependent variable is not continuous. The
duration of ofﬁce becomes a discrete ordered categorical variable. Based on an ordered
polytomous dependent variable, the ordered probit model will be appropriate. The

observed variable, Y2,, is an ordinal variable. Thus, we assume that the variable of

theoretical interest, Y,’ which is unobserved and should be interval, ﬁts a normal

90

Table l

Dummy Variables

 

Demographic Attribute

Career Attribute

 

 

Place of Birth (BPLACE)

URBAN = Urban+Capitol
[Rural+Foreign]

Educational Level (DEGREE)

COLLEGE = Above College
[Below High]

Previous Career (CAREER)

LAWPOL = Lawyer/Party Oﬁ'J
Pol. Activist

CIVSERV = Civil Service

MILITAR = Military

LIBTECH = Liberal/Business/
Own Farm/Foch.

[Laborer/Small Wage/Farmer/

Agriculture]

Long-term Foreign Experiences
(FOREIGN)

FORSTUD = Study
FORWORK = Work
FORBOTH = Both
[NONE]

Political Ideology (PID)

LIBERAL = Liberal
CENTRE = Ind/Cent
CONSERV = Conservative

Political Experience (POLEXP)

POLONE = Party or Ministry or
Congress

POLTWO = Party+Ministryl
Ministry+Cong/
Party+Cong

POLALL = Party+Mini.+Cong

Mode of Being a Leader
(ENTRY)

NCONSTI = Nonconstitutional
[constitutional]

Institutional Attributes

 

Region (REGION)

ASIA = Asia
AFRICA = Africa
MIDEAST = Middle East

EURNAME = Europe/N.Amer
LATIN = Latin America
[Oceania]

Type of Head of Government)
(LEADTYPE)

PREZ = President/Military
[Prime Mini/Gen. Sec.]

Party System (PARTYSYS)

SINGLEP = Single Party
MULTIP = Multiparty
[NO Party]

Status of Freedom
(FREESTAT)

PARTFREE = Partly Free
FULLFREE = Free
[Not Free]

Country Age (COAGECAT)

OLD = 31 + Years
[0 - 30 Years]

 

[ ] is the category that is omitted from each variable in order to avoid perfect multicollearity.

91
distribution.

The relationship between the categories of Y2, and the values of Y,’ is speciﬁed
based on the threshold values. Even though Y2, is ordinal, it can be represented as a
series of dichotomous variables. Since the dependent variable, in this test, is categorized
into three categories (1 - 3), there will be two thresholds (m, #2) to be estimated with
u, =0. The estimation for the model will be based on the log likelihood function to use
maximum likelihood method, which is concerned with picking estimates that imply the
highest likelihood of having obtained the observed sample of the dependent variable.

All independent variables in both the selection and the substantive equations are the
same. Most personal background variables and institutional characteristics are nominal
in the model. However, AGE and, some institutional variables such as HOMOGEN,
LITERACY, POPDEN and ECONFREE, are continuous. The independent variables for the
models are not difﬁcult to operationalize. Each category of the nominal exogenous
variables is considered as a dummy independent variable. The Speciﬁc Operationalization

for the categorical variables is shown as Table 1 (also, see APPENDIX A).

Interpretation of Results

Estimates for the probit model are developed by the method of maximum
likelihood. A probit analysis yields estimates that are true probabilities and which lie
in the appropriate range (Aldrich and Cnudde, 1975, p.581). Consequently, it is
possible to more accurately reﬂect the qualitative choice being made and how it has
been measured. In a probit model (selection equation), estimated coefﬁcients reﬂect

the effect of a change in an independent variable upon the inverse of the normal CDF.

92

However, the coefﬁcients should not be interpreted as straightforward as with regression
coefﬁcients (Aldrich and Cnudde, p. 580). The amount of the increase in the
probability depends upon the original probability and, thus, upon the initial values of
all the independent variables and their coefﬁcients.

An increase or decrease of one standard deviation in each independent variable
has effect on Pr(Y,, = 1), assuming that all other variables are currently at their mean
(at Z = 0.0 and probability = .5). The Sign of the coefﬁcient indicates the direction of
the change the magnitude depends upon F(Z,), and reﬂects the steepness of the CDF at
Z,. The steeper the CDF, the greater the impact of a change in the value of an
explanatory variable. Meanwhile, the estimated standard errors indicate the usual
measure of the likely variation in the estimated coefﬁcients that one anticipates to arise
from sample to sample. In terms of importance, variables with relatively large
coefﬁcients (in absolute value) have the potential to have the greatest impact on the
outcome of political survival. They can potentially have the greatest inﬂuence on the
estimated probability of whether the dependent variable takes on the value of 0 or 1.

The individual coefﬁcient estimates in a probit model measure the relationship
between the exogenous variables, X,s, and the endogenous variable, Y,. The t-score is
used for testing the H0 that a coefﬁcient ([3,) is 0. The test-statistic is deﬁned as t, =
Bkls, and is used to see whether the null hypothesis can be rejected. For the test of the
signiﬁcance of individual coefﬁcients, such as individual hypotheses testing, standard
normal tests (t-statistics) can be applied just as with OLS regression.

For the analysis of the substantive equation (ordered probit), as for the

selection equation (dichotomous probit), the standard coefﬁcient (Bl...) represents the

93

number of standard deviations of change in the dependent variable (YE), duration
indicator, brought about by a change of an standard deviation on the independent
variable. Thus, it allows one to compare the strength of different variables in the same
equation. The Y: change with the independent variables according to the sign and
magnitude of the corresponding coefﬁcients; and the magnitude of the change is
constant at all levels of the independent variables.

Instead of using an estimated R2 as a goodness-of-ﬁt measure as in the OLS, the
likelihood-ratio test statistics will be used for the probit model. While likelihood-ratio
is usually used as a test statistic for overall signiﬁcance of the model, it actually takes
on the role as a goodness-of-ﬁt measure. A test of the overall signiﬁcance of the
independent variables (H0 : B2 = [33 = = [3,, = 0) can be based on the different log

likelihood ratio. The likelihood ratio is deﬁned to be:

A = L(0')/L(0)

A’ = -210gA -2 [L’(0') - Men

-2 (LLR0 - LLR,) .

. While L°(0) (= LLR,) is the log of the likelihood function for the full model as ﬁtted
and evaluated at point 0 with all the parameters, L'(0') (= LLRO) is the log of the
likelihood function evaluated at point 0' with a constant term (intercept) only (restricted
model). A' (= -2LLR) is tested by the chi-square distribution with the number of
coefﬁcients as the degrees of freedom. -2LLR yields a measure of the statistical
signiﬁcance of the improvement over another model, but does not give any direct

indication of the model’s overall ability to explain variation in Y,. The higher the

94

likelihood, the better the estimations.

For a goodness-of-ﬁt measure, R2 is not recommended, even though it can be used
in the case of a binary dependent variable. In ordered probit analysis, R2 is only an
estimate of the true R2 because we cannot observe either the residual about the
regression plane or even the deviations of the dependent variable about its mean
(McKelvey and Zavoina, 1975, p.112). However, several pseudo-R2 have been
introduced in order to provide a measure of goodness of ﬁt.'°

According to Hagle and Mitchell ( 1992, p.774), it is better to use Aldrich and
Nelson’s measure, pseudo-R2 = -2LLR / (N - 2LLR), because of its computational ease,
lower standard errors, and higher R2s when regressed on the OLS R2. Furthermore,
even a correction for the Aldrich and Nelson pseudo-R2 is introduced. However, in this
research, any of these R23 will not be used as a measure of goodness of ﬁt because the
model itself is composed of probit models as well as ordered probit models. Although
pseudo-R2 can be applied to test a goodness of ﬁt of a probit model, it is useless in the
test of an ordered probit. There is no obvious intuitive meaning one can attach to this
measure.

It is necessary to predict the probability that the dependent variable is in each of
3 categories for any combination of the independent variables. Since both predicted and
actual values of the dependent variable are ordinal variables, it is easy to determine the

strength of association between the two. AS a result, accuracy can be evaluated either

 

'0 McKelvey-Zavoina (1976), Adlrich-Nelson (1984), and Dhrymes (1986) were compared and
evaluated in Hagle and Mitchell (l992)’s work.

95

in terms of the ﬁt between the two or in terms of the model to forecast observed Y2,
With a success table, it is also possible to compare the improvement of prediction.

For a goodness-of-ﬁt of the model, it is frequently worthwhile to report the in-
sample predictive success of the model, that is, the number of correct predictions, where
a prediction is correct when 13,25 and Y,,=1 or P,<.5 and Y,,=0 in the case of probit
model. In other words, a useful summary of the predictive ability of the model is a 2
x 2 table of the success and failure of the prediction rule. Similarly, for the ordered
probit model (substantive model), a 3 x 3 table of success and failure can be set up
based on its prediction rule.”

Though not as common as other measures of goodness-of-ﬁt, proportional
reduction error (ROE) is sometimes available. The Goodman and Kruskal (1954) A‘2
is a measure of the reduction in predictive error in one class due to knowledge of a
second class (Weisberg, 1978; Liebetrau, 1983, pp.16-24; Kay, 1991, pp.128-130). It
is a measure of a multivariate model’s improvement over‘a selected baseline, rather than
a general model’s performance. Thus, the usual baseline is the modal category.

The procedure of getting ROE is almost the same as other measures in the

crosstabulations. In the crosstabulation of nominal variables, the degree of association

 

'1 The prediction rule for the dependent variable (Y2,) with 3 categories is shown as:
P1 = Pr(Yn = 1) P2 = Pr(Yas = 2) P3 = Pr(Yza = 3)
Prediction = 1 if [(p1 2 p2) & (pl 2 p3)]
Prediction = 2 if [(p2 > pl) & (p2 2 p3)]
Prediction = 3 if [(p3 > pl) & (p3 > p2)]
’2 Reduction of Error (ROE) is calculated by the following formula:

% correctly predicted - % modal

0 =
ROE (A) 100 x 100 - % modal

 

96

is described by how well one can make predictions about the values for members on
one variable based on their values on the other variable. The measure is deﬁned as the
proportion of errors reduced made, when information on the independent variable is
used instead of ignored in making the predictions. Thus, the magnitude of measure of
association may be interpreted as the proportional reduction in error. An error consists
of a misclassiﬁcation of a member on the dependent variable. The larger the relative
reduction in error, the better the dependent variable can be predicted using the

independent variable and the stronger the association between the variables.

Descriptive Analysis

First of all, in order to ﬁnd what the similarities and differences between political
leaders as heads of government are, the deﬁnition of cleavage or heterogeneity or
fragmentation Should be introduced. Especially, in regard to their social backgrounds,
career backgrounds, and personality or skills, it is possible to compare the variation
based on homogeneity scores. Homogeneity scores based on personal background
characteristics will be compared by the groups that indicate systematic differences such
as duration, type of removal, region, type of head of government, and number of entry.

Although the sample is a nonrandom purposive sample of political leaders with
possibilities as well as limitations, it is possible to develop, consistent with canons of
contemporary social science, descriptive - historical statements - or "modal models" -
of leader traits and skills, and their social and career backgrounds. In order to make
the data from diverse cultures and time periods functionally equivalent, the information

of ethnocentric or otherwise biased materials must be minimized as much as possible.

97

The sample (882) representing the leaders who are out of ofﬁce is used for the
analyses with duration of ofﬁce and type of removal, whereas the whole sample (997)
including the present leaders, is used for the other analyses such as region, type of head
of government, or number of entry. It is based on the descriptive statistical method of
summarizing the information through crosstabulations. Some variables used in the
descriptive analysis are collapsed into different categories for different purposes.

Rae and Taylor (1970, pp.26-33) introduced how to measure the homogeneity
index. However, Kim and Patterson (1991, p.198) modiﬁed Rae’s formula for the
fragmentation score, Since different background characteristics are coded in different
numbers of categories and it is impossible to compare the scores across variables.
Especially with a data set which includes several missing values in some variables”,
a certain effort is needed to utilize available data without discarding all cases with
missing cases values. Some social background variables, such as SES and father’s
A occupation, have been omitted from the inferential models because those variables have
higher percentages of missing cases. However, they are included in the descriptive
analyses as important elements.

Considering that the cases are divided equally between 11 categories", the

 

'3 In this study, some personal variables have relatively high percentages of missing data: place of
birth (7.9%), major (2.6%), SES (35.9%), father’s occupation (40.9%), and orals kills, organimtion skills,
language skills, parents’ death/separation in early age, physical deformity, and power connection (22.3%).
Also, there are some other personal variables that have small percentages of missing data such as age
(.8%), education (1.0%), foreign experience (.9%), prison/exile experience, political activity, military
experience, previous career (.8%), experience in congress (2.1%), experience in party (2.0%), experience
in ministry (1.8%). In case of the institutional variables, homogeneity index (5.1%), party system (.2%),
and literacy (1.0%) have missing values.

" The range of F score is 0 to 1. If each individual is in a category by himself (N=n), the upper
bound for F is 1.

98

formula for Rae’s fragmentation score should be modiﬁed. The measure of
homogeneity score (H) = l - F', where F’” is the adjusted fragmentation score. For
an example, if SES of the Asian leaders (114 cases) is divided into 3 categories [low
(27), middle (23), high (64)], Fmam = 114 x (3 - 1) / (114 - 1) x 3 and F = (272 + 232
+ 642) / 1142. As a result, the homogeneity score on SES of the Asian leaders is .123.
Thus, when interpreting this score, it is not important what category has the highest or
lowest frequencies. Rather, it just indicates how similar or different the leaders in a

certain group are on a certain variable.

 

'5 F' = F/Fm, where F is Rae’s fragmentation score and Fm is the maximum score an index can
attain given its number of categories. The formulas for F and F' are as follows:

._'_
N(N-l)

PM = N(n - 1)/n(N-l)

F=1- ism-1)

CHAPTER V

Presentation and Discussion of Results

As discussed in Chapter III and Chapter IV, 984 top executives (heads of
governments) across 147 countries were selected by some manipulation and 17
propositions were introduced. A two-equation model is composed of 32 independent 4
variables (5 continuous variables and 27 dummy variables). The dependent variable,
’duration in ofﬁce’, is used according to two different perspectives. For the selection
equations, it is categorized into either ’SHORT (0-2 years)’ or ’NOT-SHORT (3 +
years)’. ’NOT-SHORT’ is coded as 1 (581 cases) and ’SHORT’ as 0 (403 cases).
However, in the substantive equations, only ’NOT-SHORT’ (581 cases) is included for
testing and analyzed as three different categories: 1 (3 - 6 years), 2 (7 - 12 years), and
3 (13 years + ).

In this chapter, ﬁrst of all, in order to test the propositions on social backgrounds,
career backgrounds, and institutional variables, the results of each reduced model will
be discussed. With the reduced models, it is predicted to see whether social, or career
backgrounds, or institutional attributes as an individual subgroup is a valid predictor of
political survival as political success. Second, an interaction model scrutinizes
interactional effect after combining all three subgroups together. Being different from
the results of the reduced models, it will present different explanation and prediction.

Although the career background model (64.7%) among the reduced models
indicates the best goodness-of-it for predicting political survival (Selection equation),

the interaction model (66.55%) shows a better ﬁt than others. Meanwhile, in case of

99

100

the substantive equation (the strength of political survival), the institutional model
(64.65%) has the highest percentage correctly predicted, but the interaction model
(64.74%) surpasses the institutional model (64.65%) in prediction by a shade of
difference. Therefore, interpretations of the results from the interaction model will be
more emphasized and examined according to each hypothesis.

Finally, at the end of this chapter, the interaction model will be tested by regional
differences. In the previous tests, the regional dummies do not Show any signiﬁcance.
Therefore, in order to see the regional effect, it is worth while dividing the sample into
different regional groups and testing each interaction model. Even though Oceanic
countries are excluded from the interpretation for the lack of enough cases, it is possible
to ﬁnd the preferences on political leaders in each region. Personal backgrounds have
greater impact on political survival in Latin America in comparison with other regions.
However, for explaining the strength of political survival, more personal background
variables have signiﬁcant impact in both the African model and the Middle Eastern
model.

According to the goodness-of-ﬁt measures for the regional models, the Middle
Eastern model (83.33% correctly predicted) explains leaders’ duration, short or long,
whereas the European model shows the lowest ﬁt (67.69%). In contrast, for the
strength of political survival, the African model (53.61%) indicates the lowest ﬁt and
the Middle Eastern (83.97%) and Latin American (87.16%) models indicate better ﬁts

than others.

101
A. Statistical Analysis of the Two-Equation Model

Based on the theoretical propositions mentioned in Chapter III, variables are
operationalized and tested for each hypothesis. The results from the OLS model, which
is based on continuous duration data do not differ much from those in the substantive
equation of the interaction model (see APPENDIX C). The independent variables that
are signiﬁcant in the OLS model are also signiﬁcant in the ordered probit model.
F reign experiences, ideology, and party system variables turn out to be different. While
both SINGLE? and MULTIP are signiﬁcant in the ordered probit, only MULTIP is
Signiﬁcant in the OLS model.

In the OLS model, R2, the goodness-of-ﬁt measure comes out to be .20
(corrected R2 is .17) which is not good enough. Except R2 statistics, there is nothing
else to evaluate a model as a whole. Thus, as far as similar results on signiﬁcance tests
can be seen later in the ordered probit model, the OLS can not be interpreted in any
other way. Therefore, this section presents the results of a two-equation model with
both the selection (dichotomous probit) and the substantive equations (polychotomous
ordered probit), which utilize the same operationalized variables as used in the OLS.
In fact, before scrutinizing the whole model (Interaction model) itself, the results from

each reduced model will be discussed step by step.

Analysis on Social Backgrounds

The results of both the selection (probit) and the substantive (ordered probit)
equations, which estimate a dependent variable, SURVIVAL, with just social background

variables, are given in Table 2. Both equations are composed of three independent

102

variables, but the substantive equation has one more exogenous variable, the hazard rate

(A,), which is calculated from selection equation in order to adjust for selection bias.

Selection Equation :

SURVIVAL = a + B,AGE + BzCOLLEGE + BJURBAN + u,

Substantive Equation :

STRENGTH = a + [3,AGE + B,C0LLEGE + B,URBAN + haz + II,

In the selection equation, whether the estimated R2 turns out to be signiﬁcant or
not, it is not used as a measure of goodness of ﬁt. This model predicts 61.17% of the
durations correctly. The model as a whole, using an analog to F-statistics, -2LLR, is
easily signiﬁcant at the .001 level with a value of 70.48. Nonetheless, actually, the
substantive equation has a better ﬁt than the selection equation since the percentage
correctly predicted (62.12 %) and -2LLR (96.28) is higher in the substantive equation.

In terms of the individual coefﬁcients, all three of the social backgrounds are
Signiﬁcant: AGE and URBAN are signiﬁcant at .05, and COLLEGE at .10. Surprisingly,
however, the directions of URBAN and COLLEGE are different from the predicted ones.
In other words, these three social background variables have signiﬁcant relationships
with whether leaders stay in ofﬁce for ’SHORT’ or ’NOT-SHORT’, but, in fact, the leaders
who had degrees above college and were born in urban areas less more likely to stay
in ofﬁce for longer periods.

The negative sign and relatively weak signiﬁcance of COLLEGE suggests that

military education is considered as a category of above college, and that usually military

103

leaders are more likely to stay in ofﬁce shorter because they are coup d’etat leaders,
transitional or care-taker leaders. Even though URBAN has a signiﬁcant effect on
whether the leader stays in ofﬁce for a short period or not, unexpectedly, the direction
of the relationship turns out to be opposite. The greater chance to be exposed to

politics and media in urban areas does not have any roles on political socialization.

Table 2 Coefﬁcient Estimates for Social Background Variables

 

 

 

 

 

Selection Equation Substantive Equation

Dependent Variable: SURVIVAL Dependent Variable: STRENGTH

Variable MLE S.E. t-score Variable MLE S.E. t-score

Constant 1.274 .247 5.167‘" Constant .017 .102 0.163

AGE -.112 .040 3029"" AGE .706 2.066 0.342

COLLEGE -.238 .131 -1.813 " COLLEGE .560 1.886 0.297

URBAN -.224 .086 -2.603"'" URBAN -7.960 20.228 -0.394
haz 5.814 1 1.965 0.486
It .764 .061 12.552‘“

% correctly predicted 61.17 % correctly predicted 62.12

-2 x LLR 70.48“" -2 x LLR 96.28"”

N 904 N 557

 

 

" signiﬁcant at .10
” signiﬁcant at .05
”" signiﬁcant at .01

With the data (581 cases) from the selection equation, the same independent
variables are tested with three categories of duration (DURATION) in order to test the
relationships between the social background variables and the strength of political
survival. The most interesting ﬁnding is that all the exogenous variables are not

signiﬁcant in the substantive equation. Hence, those three social background variables

104

really matter with whether the leaders stay in ofﬁce for a short time or not, but do not
inﬂuence the strength of political survival at all beyond this point. i

As a result, these social background variables are very important elements not
for the leaders to hold their power longer, but for the people when they elect or choose
their leaders. It suggests that after a certain period, i.e. a honeymoon period, the social
backgrounds are not considered as signiﬁcant requirements for the leaders’ maintenance
in ofﬁce. During the honeymoon period, the leaders with a peculiar blend of activism
and inexperience, prepare to build up new policies and try to mount desires for

innovation at precisely the time expertise is lacking as, a weapon for survival.

Analysis on Career Backgrounds

Selection Equation :

SURVIVAL = a + [3,LAWPOL + 8,CIVSERv + B,MILITAR + p,LIBTECH + p,I>oLoNE +
B6POLTWO + 8,I>oLALL + 8,I=oRSTUD + p,FonoRK + 8,,FoRBoTH +
[3,,NCONSTI + 8,,LIBERAL + [3,,CEN'I'RE + [3,,CoNSERv + u,

Substantive Equation :

STRENGTH = or + B,LAWPOL + BZCIVSERV + p,MILITAR + 8,LIBTECH + BSPOLONE +
8,I>oLTwo + B,POLALL + B,F0RSTUD + B,I=onoRx + 8.,F0RB0TH +
[3,,NCONSTI + 5.,LIBERAL + 8.,CENTRE + 8.,CoNSERv + haz + u,

Among the previous occupations in which the leaders were involved, the
politically related ﬁelds, such as lawyers, political activists, party ofﬁcials, as well as
civil servants (ministry jobs) have great impact on leaders’ political survival. Even
though liberal occupation, technical business, etc. are signiﬁcant at the .10 level (Table

3), military jObs are not crucial to decide their survival at all. However, contrary to the

105

theoretical expectations, the relationships between politically related jobs (LAWPOL and
CIVSERV) and political survival are negative.

Political experiences, whether they had experience only in party, ministry, or
congress, are necessary requirements for the leaders to stay in ofﬁce longer. The
degree of political experience or its experience in some special places, in fact, does not
matter, but the leaders should have at least some political experience. In case of foreign
experiences, working in foreign countries itself does not have any impact on political
survival. However, studying in foreign countries has signiﬁcant role in deciding
whether or not a leader stays in ofﬁce for a shorter period.

According to socialization theory, political socialization is more effective at
schools than at work areas. Especially, most jobs in foreign countries are military
attache, diplomatic mission, business, and so on. Thus, foreign experiences in those
jobs would not have limited impact on new political socialization. However, when
leaders were studying abroad, they would acquire certain political perspectives and
develop their own views on politics. For example (see Figure 7), high percentages of
leaders except in North American and European countries had foreign experience. This
trend indicates that leaders from underdeveloped or developing countries would go to
foreign countries to obtain advanced education and they have more opportunities to see
all different aspects of politics in countries other than their own.

As for the hypothesis on entry mode (NCONSTI), it is expected that leaders by
nonconstitutional entry are more likely to stay shorter. However, the null hypothesis
is not rejected, Since the t-score of NCONSTI is lower than its critical value, while means

that nonconstitutional entry does not have any relationship with political survival at all.

106

Table 3 Coefﬁcient Estimates for Career Background Variables

 

Selection Equation

Substantive Equation

 

Dependent Variable: SURVIVAL

Dependent Variable: STRENGTH

 

 

 

Variable MLE S.E. t-score Variable MLE S.E. t-score
Constant -0.335 0.347 -0.968 Constant 0.473 3.707 0.127
LAWPOL -0.598 0.246 -2.433” LAWPOL -1.171 2.494 -0.470
CIVSERV -0.600 0.267 2244" CIVSERV -1.459 2.504 -0.583
MILITAR -0.217 0.268 -0.811 MILITAR -0.608 0.987 -0.616
LIBTECH -0.445 . 0.243 -1.828"‘ LIBTECH -1.004 1.893 -0.530
POLONE 0.5 13 0.247 2.077" "' POLONE 0.509 1.725 0.295
POLTWO 0.643 0.269 2.390" POLTWO 0.504 2.260 0.223
POLALL 0.609 0.271 2.246" POLALL 0.408 2.118 0.193
FORSTUD 0.243 0.103 2.351 " FORSTUD 0.601 0.964 0.623
FORWORK 0.189 0.122 1 .554 FORWORK 0.323 0.769 0.420
FORBOTH 0.328 0.139 _ 2.350” FORBOTH 0.485 1.327 0.366
NCONSTI -0.101 0.131 -0.773 NCONST'I 0.320 0.428 0.748
LIBERAL 0.381 0.178 2.142" ‘ LIBERAL 0.747 1.490 0.501
CENTRE -0.046 0.233 -0.197 CENTRE 0.372 0.341 1.092
CONSERV 0.505 0.173 2921"“ CONSERV 0.939 1.991 0.472
haz -1.328 5.715 -0.232
p 0.734 0.058 12.636”*
% correctly predicted 64.70 % correctly predicted 61.82
-2 x LLR 9840*” -2 x LLR 7334*"
N 949 N 571

 

 

" signiﬁcant at .10
” signiﬁcant at .05
“‘ signiﬁcant at .01

107

With the exception of most military leaders, almost all leaders should represent
their political ideologies in their elections or in their decision-making processes. The
results Show that the leaders whose ideologies are liberal or conservative are more likely
to stay in ofﬁce longer. The direction of the relationship between LIBERAL and political
survival is positive although it would be expected to be negative.

Just as for the social background variables, all career background variables do
not have any relationships with political survival in the substantive equation, supporting
the previous argument that the same career backgrounds as personal attributes are
signiﬁcant in explaining political survival, i.e. , SHORT or NOTSHORT rather than the
strength of political survival. In the case of career backgrounds being different from the
social background variables, the selection equation (64.7 % correctly predicted and with -
2LLR (98.40)) shows a better ﬁt than the substantive equation (61.82% correctly
predicted and with -2LLR (73.34)). While the selection equation explains better in
career backgrounds than in social backgrounds, the substantive equation has a better ﬁt
in social backgrounds. From these results, it is apparent that career backgrounds are
more important in deciding whether a leader survives beyond a short period, but they do

not affect how long the person stays in ofﬁce after that initial test period.

siso I ti i 1A 'b

Selection Equation:

SURVIVAL = a + B,OLD + 13sz + 8,SINGLEI> + B,MULTIP + BSHOMOGEN + B,PoI>DEN
+ B,LITERACY + 8,I>ARTFREE + 8,FULLFREE + 8,,EC0NFREE + [3,,ASIA +
8,,AFRICA + B,,MIDEAST + BMEURNAME + p,,LATIN + u,

l 08
Substantive Equation:

STRENGTH = a + B,OLD + mm + I3,SINGLEI> + [34MULTIP + 8,H0M0GEN + [36POPDEN
+ B,LITERACY + B,PARTFREE + B,FULLEREE + 8.,EC0NFREE + [3,,ASIA +
8.,AFRICA + B,,MIDEAST + [3,,EURNAME + B,,LATIN + haz + It,

Again, the independent variables that turn out to be signiﬁcant in the selection
equation become insigniﬁcant in the substantive equation. Like personal variables
(social and career), institutional variables do not explain the strength of political
survival, although some of them are signiﬁcant in deciding whether leaders stay in
ofﬁce for a short period or not.

Country age (OLD) is an important element for political survival and its direction
of the relationship is expected to be positive, but it comes out negative. Thus, leaders
in older countries are more likely to stay only a short time in ofﬁce. As far as political
stability and tranquility are concerned, the leader in old countries can maintain his
power longer.

Nonetheless, while European and North American countries have stable liberal
parliamentary democracies, and Latin American countries often have had military
regimes, those political sytems which lead to frequent leadership succession.
Meanwhile, African countries and several Asian countries are young, having only been
independent since the 19605. If a leader was a founder or a ﬁrst leader in those
countries, they were often top leaders for a long time until their new political regimes
become stable or developed in a certain degree.

Presidential system (PREZ) supports leaders to stay longer than parliamentary
systems (and general secretaries). Presidential systems guarantee constitutional terms

(4-6 yrs). Under prime minister systems, top executives confront elections whenever

109

Table 4 Coefﬁcient Estimates for Institutional Attributes

 

 

 

 

 

Selection Equation Substantive Equation
Dependent Variable: SURVIVAL Dependent Variable: STRENGTH
Variable MLE S.E. t-score Variable MLE S.E. t—score
Constant -0.195 0.381 -0.51 1 Constant 2.331 2.756 0.846
OLD -0.264 0.124 -2.134” OLD -0.989 0.772 -1.280
PREZ 0.589 0.124 4.763” PREZ 1.369 1.755 0.780
SINGLEP 0.388 0.186 2.089” SINGLEP 1.233 1.187 1.039
MULTIP -0.055 0.167 -0.327 MULTIP -0.344 0.253 -1.359
HOMOGEN -7.5e-05 2.3 e-04 -0.33 1 HOMOGEN -2.7e—4 3.1e-4 -0.858
POPDEN -7.8e-05 2.8e-04 -0.282 POPDEN 2.2e-4 3.8e-4 0.596
LITERACY 0.008 0.024 3.246‘” LITERACY 0.020 0.230 0.872
PARTFREE 0.139 0.157 0.885 PARTFREE 0.228 0.459 0.497
FULLFREE 0.737 0.209 3530“" FULLFREE 1.127 2.189 0.515
ECONFREE -0. l 35 0.060 -2.245’ "‘ ECONF REE _ -0.279 0.399 -0.700
ASIA 0.041 0.265 0.156 ASIA -0.092 0.332 -0.276
AFRICA 0.037 0.276 0.133 AFRICA 0.086 0.327 0.262
MIDEAST -0.166 0.269 -0.617 MIDEAST -0.583 0.593 -0.983
EURNAME -0.278 0.235 -l.182 EURNAME -0.630 0.892 -0.706
LATIN -0.23 1 0.251 -0.923 LATIN -1 .284 0.774 -1.658‘
haz -3.245 4.191 -0.774
p 0.843 0.666 12.649‘“
% correctly predicted 62.62 % correctly predicted 64.65
-2 x LLR 92.70"” -2 x LLR 165.44"“
N 947 N 563

 

 

" signiﬁcant at .10
” signiﬁcant at .05
"* signiﬁcant at .01

llO

parliamentary members vote no conﬁdence, sometimes as frequently as every other year.
Even though general secretaries in communist countries are similar to prime ministers
in terms of the selection process, they usually stay in ofﬁce until their deaths.

SINGLEP is signiﬁcant at the .05 level and it supports the relationship between
party system and political survival that we expected earlier, whereas MULTIP is not
signiﬁcant. Single-party systems are stable only in the formal sense that a single party
remains in power for decades. The impact of parties is considered as strengthening the
link between rulers and the population. However, although it is believed that some
countries are ofﬁcially based on multiparty systems, only one party always takes the
initiative in politics. In this case, the party system should be considered as a
monopolized one-party system in reality. In order to have better outcomes from this
model, more careful inspections of party systems are required.

One of the most unexpected results is generated from HOMOGEN and POPDEN,
which fail to support our expectations about relationship between political stability and
leadership durability. The indicators of govemability, the homogeneity index and the
population density were expected to be signiﬁcant on political stability, and, ultimately,
on the leader’s durability. Unexpectedly, however, there are no relationships between
political survival and any of them.

Considering on the freedom status, not only in political and civil rights but also
in economic ﬁeld, FULLFREE and ECONFREE are proven to be signiﬁcant. While freer
societies, in the terms of political and civil rights, allow their leaders to stay longer, high
economic freedom (ECONFREE) has a role against the leaders’ longevity. If ofﬁcials

employ force, that fact directly reﬂects a loss of power regardless of any legal

111

justiﬁcation that those ofﬁcials are able to invoke. As a result, those who rely primarily
on force are likely to be politically unsuccessful, whether or not they assumed ofﬁce
through a military coup. Successful regimes, military or otherwise, are those that are
capable of exercising power. The repeated invocation of force is evidence of a loss of
power and it accrues to the level of freedom guaranteed.

Meanwhile, the economy with high level of freedom is inherently unstable;
economic growth is constantly disrupted by crises. In a country with low economic
freedom, intervention in the economy and other Spheres grows and draws attention to the
issues of choice, planning and control. 80, the ’hand of the state’ is more visible and
intelligible than ’the invisible hand’ of the market. Consequently, as long as the
governments take on responsibility for more and more areas of the economy,
governments can avoid economic and political crises and ﬁnally leaders would keep their
ofﬁces longer.

Regional difference as a control variable, in effect, does not show any signiﬁcance
in explaining political survival. However, in the substantive equation, only LATIN is
signiﬁcant at the .10 level. Furthermore, it has a negative relationship to the strength
of survival, indicating that the level of longevity can be explained better in Latin
America, and that the Latin American leaders are more likely to stay in ofﬁce shorter.
Apparently, in Latin America, we see quite a few leaders who just stayed in ofﬁce as
coup d’etat leaders for a short period or were themselves overthrown again by the
military. Also, by constitution, several countries in this area do not allow the same
leaders to be elected in a consecutive term or their tenure is limited to a certain period

(see note 6).

112

In general, the statistics for testing a whole model with the institutional attributes
Show the same tendencies as shown in the previous social and career background models.
The independent variables that are signiﬁcant in the selection equation become
insigniﬁcant in the substantive equation (Table 4). Both the selection and substantive
equations, as a whole, are signiﬁcant at the .01 level. However, the substantive equation
(64.65 %) has a higher proportion of prediction than the selection equation (62.62%).
The institutional model itself explains the strength of political survival better than

political survival.

Comparisons of the Reduced Models on Political Survival

Analyses of the success tables generated from the results are presented only for
explaining political survival. In principle, those success tables of the strength of political
survival (substantive equations) should be 3 x 3 tables, but the results show 3 x 2 tables
because the models do not predict the second category (7 - 12 years). Thus, it is not
reasonable to compare the predictions for the substantive equations. Also, in order to
use the statistics, i.e. , ROE (Reduction of Error), it is better to focus on 2 x 2 tables,
that is, the success tables in the selection equations.

In Table 5, it is possible to see how the composition of the success of prediction
is composed of. In the case of the social background model, the category of SHORT has
a big reduction, 31.8% (38.4% - 6.6%) and NOTSHORT has a big improvement, 33.6%
(61.6% -r 95.2%). In the case of the career background model, the SHORT category
shows a 13.9% of reduction of prediction, while NOTSHORT is improved by 30.2%.

Moreover, the institutional model indicates an 11.1% reduction in predicting SHORT, and

PREDICT

PREDICT

PREDICT

113

Social Backgrounds

Table 5 Success of Prediction in Political Survival

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ACTUAL
0 1 Row
SHORT NOT-SHORT TOTAL
0 23 27 so
SHORT 6.6 4.8 5.5
1 324 530 854
NOT SHORT 93.4 95.2 94.5
COLUMN 347 557 904
TOTAL 38.4 61.6 100.0
Career Backgrounds
ACTUAL
0 1 Row
SHORT NOT-SHORT TOTAL
0 98 55 153
SHORT 25.9 9.6 16.1
1 280 516 796
NOT-SHORT 74.1 90.4 83.9
COLUMN 378 571 949
TOTAL 39.8 60.2 100.0
Institutional Attributes
ACTUAL
0 1 Row
SHORT NOT-SHORT TOTAL
0 113 83 196
SHORT 29.4 14.7 20.7
1 271 480 751
NOT-SHORT 70.6 85.3 79.3
COLUMN 384 563 947
TOTAL 40.5 59.5 100.0

114

a 25.8% improvement in predicting NOTSHORT.

As can be seen in each reduced table, and in Table 8, the career background
model represents the best goodness-of-ﬁt among the reduced models. Its proportion
correctly predicted is 64.7% and ROE is 11.37. In other words, in order to explain
political survival itself, it is better to use the career background model. However, in
terms of only the percentage of correctly predicted, the institutional attribute model
(64.65%) can predict the strength of political survival better than the other two models
social background (62.12%) and career background (61.82%). Nonetheless, all two-
equation reduced models have a common tendency: some variables that are signiﬁcant

in the selection equation lose their inﬂuence in the substantive equation.

Analysis on Interaction Model

Selection Equation:

SURVIVAL = or + B,AGE + B,COLLEGE + 8,URBAN + 8,LAWI>OL + B,CIVSERV + 8,MILITAR
+ 8,LIBTECH + 8,POLONE + [39POLTWO + 8,,POLALL + [3,,FORSTUD +
8,,FORWORK + 8.,FORBOTH + [3,,NCONSTI + 8,,LIBERAL + [3,,CENTRE +
8,,CONSERV + B,,OLD + [3,,PREZ + [3,,SINGLEP + 0,,MULT'IP + 8,,HOMOGEN +
[3,,POPDEN + [1,,LITERACY + 9,,PARTFREE + 8,,FULLFREE + 8,,ECONFREE +
[3,,ASIA + 8,,AERICA + 8,,MIDEAST + 8,.EURNAME + 8,,LATIN + p,

Substantive Equation:

STRENGTH = a + [3,AGE + BZCOLLEGE + p,URBAN + p,LAWPOL + BSCIVSERV + p,MILITAR
+ B,LIBTECH + B,POLONE + p,POLTwo + [3,0POLALL + B,,FORSTUD +
BUFORWORK + 8,,FORBOTH + 5,,NCONS'IT + [3,,LIBERAL + 8,,CENTRE +
13,,CONSERV + 8,,OLD + 8,,PREz + 8,,erGLEP + 8,,MULTIP + [1,,HOMOGEN
+ 8,,POPDEN + 8,,LITERACY + [3,,PARTFREE + 8,,FULLFREE + p,,ECONFREE
+ 8,,ASIA + 8,,AERICA + [3,,MIDEAST + [3,,EURNAME + 8,,LATIN + haz + p,

In the case of the selection equation (Table 6), the results Show that AGE,

115

URBAN, CONSERV, PREZ, LITERACY, and ECONFREE are signiﬁcant at the .05 level, and
that COLLEGE, POLALL, LAWPOL, FORSTUD, FORBOTH, and PARTFREE are signiﬁcant at
the .10 level. Except NCONSTI, OLD, HOMOGEN, POPDEN, and regional dummies, at least
one of the dummies in each variable turns out to be signiﬁcant, supporting the
relationships between these variables and political survival. As seen in the previous
three reduced models, the directions of some variables are opposite to those expected
(COLLEGE, URBAN, LAWPOL, and ECONFREE).

Some differences will now be discussed based on the comparisons between
selection and substantive equations, as well as between the reduced models and the
interaction model. First of all, in the case of the selection equation, in comparison with
the reduced models, the signiﬁcance of CIVSERV, LIBTECH, POLONE, POLTWO, LIBERAL,
OLD, and SINGLEP had disappeared. However, as shown in Table 8, the proportion
correctly predicted (66.6%) is higher than those (61.17% for social backgrounds,
64.70% for career backgrounds, and 62.62% for institutional attributes) in each reduced
model.

Secondly, in the reduced substantive equations, all signiﬁcant variables in the
selected equations became insigniﬁcant in the substantive equations. Meanwhile, some
variables stay the same and even new ones such as CIVSERV, LIBTECH, FORWORK, OLD,
SINGLEP, and MULTIP become Signiﬁcant in the substantive equation of the interaction
model. Instead, the relationships of POLALL, PARTFREE, and ECONFREE with the strength
of political survival have disappeared in the substantive equation.

All social background variables, AGE, COLLEGE, and URBAN have great impact

on both political survival and the strength of political survival, indicating the following

116

Table 6 Coefﬁcient Estimates for Interaction Model

 

Selection equation

Substantive equation

 

Dependent Variable: SURVIVAL

Dependent Variable: STRENGTH

 

 

 

Variable MLE S.E. t-score Variable MLE S.E. t-score
Constant 0.300 0.598 0.503 Constant 6.005 2.703 2.221"
AGE -0.014 0.005 2915‘” AGE -0.071 0.026 -2.7l8”
COLLEGE -0.304 0.157 -1.935“' COLLEGE -1.050 0.570 -1.842"'
URBAN -0.194 0.010 -2.004” URBAN -0.791 0.355 -2.228”
LAWPOL -0.524 0.284 -1 .845“ LAWPOL -1.822 0.992 -1 .838‘
CIVSERV -0.459 0.307 -1.495 CIVSERV -2.226 0.925 -2.407" "' "
MILITAR -0.063 0.309 -0.205 MILITAR -0.249 0.321 -0.777
LIBTECH -0.364 0.278 -1.308 LIBTECH -1.453 0.722 -2.01 1”
POLONE 0.369 0.292 1.264 POLONE 0.985 0.710 1.387
POLTWO 0.499 0.309 1.614 POLTWO 1.244 0.921 1.350
POLALL 0.591 0.312 1.894‘ POLALL 1.515 1.066 1.421
F ORSTUD 0.232 0.120 1.930" FORSTUD 1.085 0.443 2.449‘ ‘
FORWORK 0.083 0.136 0.613 FORWORK 0.484 0.237 2.037‘ ‘
FORBOTH 0.275 0.154 1 .783 ‘ FORBOTH 1.077 0.520 2.072“
NCONSTI -0.151 0.156 -0.965 NCONSTI -0.431 0.3 32 -1.296
LIBERAL 0.270 0.209 1.292 LIBERAL 0.826 0.554 1.491
CENTRE -0.068 0.266 -0.255 CENTRE 0.187 0.396 0.471
CONSERV 0.477 0.200 2.384” CONSERV 1.768 0.900 1.963"
OLD -0.2 10 0. 135 -1 .549 OLD -1 .003 0.402 -2.496" "'
PREZ 0.694 0.138 5041*” PREZ 2.194 1.265 1.735"
SINGLEP 0.083 0.243 0.341 SINGLEP 0.865 0.352 2.457"
MULTIP -0.260 0.222 - 1 . 169 MULTIP -0.933 0.547 -1.707"
HOMOGEN 9.7e-05 2.7e-04 0.356 HOMOGEN 1 .6e-04 3 .5e-04 0.466
POPDEN -6.5e-06 3 .0e-04 -0.022 POPDEN 4.2e-04 3.5e-04 1 .205
LITERACY 0.007 0.003 2.659" LITERACY 0.027 0.013 2.025“I
PARTFREE 0.073 0.1 77 0.41 1 PARTFREE 0.305 0.246 1 .243
EU LLF REE 0.73 1 0.242 3.027 " "‘ * FULLFREE 1 .956 1.367 1.43 1
ECONFREE -0.115 0.696 -1.654" ECONFREE -0.366 0.225 -1.622
ASIA 0.31 1 0.295 1.054 ASIA 0.774 0.632 1.225
AFRICA 0.105 0.306 0.343 AFRICA 0.152 0.376 0.404
MIDEAST -0.019 0.294 -0.064 MIDEAST -0.342 0.346 -0.988
EURNAME 0.014 0.257 0.055 EURNAME 0.101 0.300 0.338
LATIN 0.174 0.281 0.617 LATIN -0.374 0.428 ‘ -0.874
haz -4.258 2.503 -1.701"
p 0.905 0.073 12.441 ""
% correctly predicted 66.55 % correctly predicted 64.74
-2 x LLR 165.20“" -2 x LLR 207.52‘"
N 861 N 536

 

 

" signiﬁcant at .10
” signiﬁcant at .05
""' signiﬁcant at .01

117

relationships:
1. The younger the leaders are upon entering Ofﬁce, the longer they stay in
ofﬁce.

2. Leaders with more educations (above college) are less likely to stay in ofﬁce
longer.

3. Leaders who were born and raised in urban areas tend to stay in ofﬁce

shorter than those from rural areas.

Among the career background variables in the selection equation, only entry
mode (NCONSTI) does not have any inﬂuence on explaining the relationship with
political survival, whereas, in accounting for the strength Of political survival, all
political experiences as well as entry mode, regardless of what offices they had been
in previously, do not explain the strength of political survival at all. On the contrary,
foreign experiences strongly account for the relationships with the strength of political
survival.

Career background variables in the interaction model indicate somewhat different
results. Actually, the results represents more obvious explanations from a theoretical
perspective. Among the previous occupations, only LAWPOL is signiﬁcant and, among
the political experiences, only POLALL. The most interesting ﬁnding concerns leaders’
ideology. Whether a leader is liberal or conservative has a great impact on the leader’s
political survival in the reduced model. However, in the interaction model, the outcome
really supports the proposition that leftist (Liberal) parties tend to be more willing to
change priorities than rightist ones, primarily because change is a more legitimate

enterprise to the leﬁ than it is to the right. The relationships of career backgrounds to

118

political survival are explained as follows:

4. Except military professions, leaders with politically related occupations such
as political activists, lawyers, party ofﬁcials, civil servants, as well as even
liberal and technical professions and business, are less likely tO stay in ofﬁce
longer. Meanwhile, in explaining political survival itself, only leaders who
were lawyers or politicians before, are less likely to stay in ofﬁce longer.

5. While a part of political experiences (POLALL) explains political survival,
there are no relationships between political experiences and the strength of
political survival. Especially, foreign study is very important for political
survival.

6. Whatever the purpose of foreign experience is, they have great impact in
accounting for the degree of political survival.

7. Entry mode does not matter with political survival or the strength of political
survival.

8. With regard to the political ideology, conservative leaders are more likely .

to stay in ofﬁce longer.

In terms of the coefﬁcients for the institutional characteristics, it is apparent that
there are several changes from the selection equation to the substantive equation.
Country age in the reduced model (Institutional model) is signiﬁcant, but, in the
interaction model, it does not account for its relationship with political survival (t-score
(=-1.549) is not large for being signiﬁcant). However, it shows a strong signiﬁcance
in accounting for the strength Of political survival. Interestingly, the type of head of
government (PREZ) is very important in determining whether or not a leader stays in
Ofﬁce for a short period. However, while still signiﬁcant at the .10 level, it is not as
strong as it is in explaining the length Of his survival after the initial period.

Party system does not give an explanation of political survival itself (short or

long), whereas it clearly proves the relationships with the strength of survival. With

119

respect to the Freedom index as a combination of political rights and civil rights, the

leaders in fully free societies (FULLFREE) stay in ofﬁce longer, but the strength of

survival is not explained by it at all. In contrast, in countries where the governments

guarantee more economic freedom, the leaders are less likely to stay in ofﬁce longer.

In the model for the strength of survival, even though its t-score is almost close to the

signiﬁcance level, it is not enough to explain the relationship. Finally, it is interesting

to speculate that regional difference does not expound political survival itself as well

as the degree of durability. As a result, the interaction model shows the following

outcomes:

10.

11.

12.

13.

14.

15.

16.

Leaders in Old countries are less likely to stay in ofﬁce longer, whereas
county age does not matter much with respect to political survival itself.

The type of head of government is much more important in deciding
political survival rather than the strength of political survival, although it
still accounts for the strength.

Party system is not related to political survival, but it has a great inﬂuence
on the strength of survival. Leaders from single party systems stay in
power longer, whereas those from multiparty systems stay in ofﬁce shorter.

There is no relationship between the degree of homogeneity of society
(HOMOGEN) and political survival or the strength of survival.

Population density (POPDEN) does not matter for political survival or the
degree of political survival.

In terms Of LITERACY, leaders from societies with high levels - used as an
indicator of modernization - are more likely to stay in Ofﬁce longer.

Freedom status as a combination of political and civil rights has a positive
inﬂuence on political survival but no inﬂuence on the degree of durability.

Great economic freedom inﬂuences whether leaders stay in Ofﬁce short or
not, but does not matter with the degree of their longevity beyond the initial
trial period. The less economic freedom, the longer the leader stays.

PREDICT

120

Table 7 Success of Prediction in Interaction Model

Political Survival (Selection Equation)

 

 

 

 

 

 

 

 

ACTUAL
0 1 ROW
SHORT NOT-SHORT TOTAL
0 122 85 207
SHORT 37.5 15.9 24.0
1 203 451 654
NOT-SHORT 62.5 84.1 76.0
COLUMN 325 536 861
TOTAL 37.7 62.3 100.0

Strength of Political Survival (Substantive Equation)

 

 

 

 

 

 

 

 

 

 

ACTUAL
1 2 3 Row
(3 - 6) (7 - 12) (13+) TOTAL
1 298 93 41 432
(3 - 6) 92.3 75.6 45.6 80.6
2 7 8 8 23
PREDICT (7 - 12) 2.2 6.5 8.9 4.3
3 18 22 41 81
(13+) 5.6 17.9 45.6 15.1
COLUMN 323 123 90 536
TOTAL 60.3 22.9 16.8 100.0

121

17. In the case of regional differences, each region does not show any
signiﬁcant relationships with both political survival and the degree of
survival.

In the substantive equation with the selected sample (563), some variables gain
signiﬁcance and some lose it. While CIVSERV, LIBTECH, OLD, SINGLEP, and MULTIP are
shown to be signiﬁcant, POLALL, FULLFREE, and ECONFREE become insigniﬁcant in the
substantive equation. In addition, although PREZ is signiﬁcant at the OI = .10 level, the
test level gets weaker (.001 —) .10). Overall, the selection equation obtains 66.6%
correctly predicted, and the substantive equation predicts 64.7% correctly. Generally,
the two-equation interaction model has a pretty good performance in explaining the
relationships between exogenous variables and political survival on the one hand,
between the exogenous variables and the strength of political survival on the other hand.
Therefore, the results of the interaction model are best for interpreting the main
substantive ﬁndings of our study.

As shown in Table 7, among the correctly predicted cases (66.6%), the
interaction model performed better at predicting ’SHORT (0)’ outcomes. The model
predicted or correctly classiﬁed 84.1% of the ’NOT-SHORT (I)’ leaders and 37.5% Of the
SHORT leaders. While the SHORT category does not have a big improvement (2%
reduction), NOT-SHORT has 21.8% (62.3% —) 84.1%) of improvement. In the case Of
the substantive equation, all three categories show different aspects: category 1 (3 - 6
yrs) had 32% of improvement and category 3 (13 + yrs) does 28.8% of improvement,
whereas category 2 (7 - 12 yrs) has rather 16.4% of reduction. All together, the

proportion of the correctly predicted cases is 64.7%. The results indicate that the model

122

might be more reasonable to predict ’SHORT’ or ’NOT-SHORT’ than to predict the degree
Of political survival. Nonetheless, as shown in Table 8, the interaction model (66.55%
correctly predicted and ROE = 11.39) was the best predictor among all the models,

even though its ROE is a little bit higher than career backgrounds (ROE = 11.37).

Table 8 Goodness-of-F it Of Selection Equations

 

 

Model % Modal % Correctly ROE
Predicted

Social Background 61.62% 61.17% -1.17

Career Background 60.17% 64.70% 11.37

Institutional Attn'butes 59.45% 62.62% 7.82

Interaction 62.25% 66.55% 1 1.39

 

123
B. Analyses of Regional Models

As regional dummies, none of them show any signiﬁcant impact on political
survival or the strength of survival in both the reduced models and the interaction
model. It is worth breaking the interaction model down region by region in order to
see better regional or cultural effect. As seen in Table 9, the results provide different
explanations of political survival and the strength of survival. According to the logic
of the European mind, every territory should fall under some sovereignty, and all people
in the same geographic location should have a common loyalty and the same legal
Obligation (Pye, 1985, p.86).

If the data is analyzed by looking at different regions of the world, it can be
found that different groups of countries show very different patterns (see Table 9).
Most of the quantitative explanatory research compasses all or most countries
simultaneously. We offer a note of caution about interpreting these results too
critically, especially if the researcher has not broken the countries down into types or
regional clusters.

In Asia, among the personal characteristics, only the educational level is important
in explaining whether or not a leader stays in ofﬁce for a short period. 80, especially,
the leader with above college degree is more likely to stay in ofﬁce for 0 - 2 years.
While the leader in a highly homogeneous society and economically free society is less
likely to survive for long, the leader in a society with high literacy rate stays in Ofﬁce
longer. In developing countries, such as Asian and Latin American countries, capitalism
is equally unsuited to the goals of the revolution since it sows the seeds Of conﬂict and

antagonism.

124

Table 9 Signiﬁcant Variables in Regional Models

 

 

 

 

Region Selection Equations Substantive Equations
Asia URBAN(‘), LITERACY(‘), ECONFREE(‘) FULLFREE (0)
HOMOGENP‘)
Africa LIBERAu“), CONSERV(”), AGE(”),URBAN(“),LAWPOL(”),CIVSERV(”)
ECONFREE(“), M1LITAR(”), LIBTECH”), POLT'WO(‘),
POLALw), FORWORK(”), FORBOTH(‘),
LIBERAu"), CONSERw"), 0LD(”),
PRFZP‘), SINGLEPP‘), POPDENv),
LITERACY(”). HOMOGEN”), FULLFREE“);
ECONFREEP‘)
Middle East AGE("), CONSERV(‘), SINGLEP(”), AGE(‘), COLLEGE(’),URBAN(‘), LAWPOL(‘),

MULTTP(“), HOMOGENP‘).

M1LlTAR(‘), LIBTECH(‘), FORSTUD(“),
OIJX“),SII~1GLEP(‘),POPDEN(‘),PARTFREE(‘),
FULLFREE('), LITERACY?)

 

 

 

Europe/N. America URBAN(‘), FORBOTH”), PREzv) MULTIP(‘)
CONSERV(")

Latin America AGE(“), POLTWO(‘), POLALw), PARTFREEv)
FORSTUDP‘), FORBOTHP), NCONSTIP)
HOMOGENv), ECONFREEv), POPDENv)

Oceania NONE NONE

 

‘ signiﬁcant at .10
" signiﬁcant at .05

125
Thus, the higher the economic freedom, the shorter the leader’s duration. They need an

appropriate path where all elements participating in the process of production can meet,
capable Of supplying society with all the energies enabling it to rebuﬂd its life on the
basis of a carefully studied and comprehensive plan. These are very contradictory
outcomes that, theoretically, we did not expect. Most of the countries in South and
Southeast Asia which gained political independence since 1945, have in common not only
political orders but also economic and social orders which have been stripped of their
previous leadership. Independence has necessarily meant a change in political and
governmental leaders. Unlike its Western counterpart, some factors that are main causes
for democracy and modernization may operate as obstacles to economic development.
Nonetheless, even those relationships disappear in the model for the strength Of political
survival, only FULLFREE is signiﬁcant at the .10 level.

In the African countries, whether a leader’s ideology is liberal or conservative,
he should identify his ideology in order to keep his ofﬁce longer. Also, in a society with
a high level of economic freedom, a leader is more likely to survive long. In the
substantive equation for testing the relationships between each independent variable and
the strength of political survival, most null hypotheses are rejected at the .10 or .05
levels except COLLEGE, FORSTUD, N CONSTI, CENTRE, MULTIP, and PARTFREE, although the
directions of the relationship of those variables are different from the expected ones.

The outcome indicates that a leader with foreign experience or more political
experience has a tendency for staying in ofﬁce shorter, and that all dummy career
backgrounds in the model have a negative relationships with strength of political survival.

It is a quite unusual result that has not been found in any other regions. Since achieving

126

independence after World War H, most of the countries have experienced at least one
coup d’etat or one civil war, resulting in profound instability. However, the results from
the regional models (Africa and the Middle East) conﬁrm the proposition about country
age, such that the leaders in old countries are more likely to stay longer in ofﬁce.

The most distinctive ﬁnding from the Middle Eastern model is that, for the ﬁrst
time, COLLEGE, turns out to be signiﬁcant at the .10 level, showing that the more
educated (above college education), the longer the duration. Another encouraging result
is that a leader in a more homogeneous society stays in ofﬁce long and longer. As
similar as the trends in African leaders, more experience in the political arenas as well
as foreign experience are not the supportive qualities for maintaining their power.

POPDEN has never been signiﬁcant in any other models except the substantive
equations in Africa and the Middle East. Many of the underdeveloped or developing
areas have become the most densely populated regions on earth because of the
introduction of modern medical and health-related practices into the societies. Thus,
the region becomes a main source of change and social dislocation. It supports the
proposition on govemability such that a leader in a sparse area is not easily threatened
by political instability and thus is able to acquire efﬁciency and equity. The size of a
country will have a direct bearing upon the kind of power structures that develop there.
To be more speciﬁc, the smaller the country, the more concentrated its power structure
is likely to be.

Next, in case of the European and North American countries, FORBOTH and
CONSERV have signiﬁcant impacts on explaining political survival. Western countries

are characterized by a high degree of internalized control and a low degree of

127

extemalized control. Therefore, it is not expected that foreign experiences are important
for the leaders in this region because most of the countries are already politically and
economically developed enough. URBAN and PREZ are shown to be signiﬁcant, even
at the .10 level. Urban born, as an indicator of political socialization, has a negative
relationship with political survival, whereas leaders as presidents are more likely to have
relatively long duration.

As to the strength Of political survival, being different from Aﬁ'ica or the Middle
East, almost all exogenous variables except MULTIP do not have any inﬂuence on it.
A multiparty system has a negative effect on how much longer a leader can stay in
ofﬁce. In this region, like in Asia, AGE is not explained by a life expectancy theory,
but rather by a gerontocracy thesis. The seniority system effect indicates that power,
prestige, and wisdom accrue to age. Seniority is a major repository of the experience
and knowledge of society and is treated with the utmost respect and deference.

The most distinctive difference in Europe and North America, HOMOGEN does
not explain anything on political survival or its strength. Heterogeneity in society is an
inescapable facet Of politics within as well as between countries; internal conﬂict and
instability and external conﬂict and war Oﬁen ﬂow from ethnic, linguistic, and religious
tensions and processes of mobilization. Nonetheless, in this region, heterogeneity will
recede in importance as the modernization process accelerates. The example for this
phenomenon that accounts for an increasing share of non-revolutionary violence and
non-violent movements are: the Basques in Spain, the Flemings and the Walloons in
Belgium, the Bretons in France, the British Celtic in Scotland, North Ireland and Wales,

and the French in Quebec.

128

In Latin America, the more political experiences and foreign experiences, the
longer the duration. In comparison with other regions, more personal variables have
an inﬂuence on explaining political survival (SHORT or NOTSHORT). The social and
cultural matrix within which Latin American leaders operate at present is such that
effective and representative popular democracy is, with really few exceptions, not a
feasible alternative (Tannenbaum, 1966, p.144). Thus, the personalist state tends to be
governed by a single caudillO-type ruler.

The archetypical examples Of personalist states include Cuba’s Batista,
Venezuela’s Jimenez, the Dominican Republic’s Trujillo, Paraguay’s Stroessner, and the
Somoza’s Nicaragua. For the institutional type, the state is essentially an instrument
Of class domination and often the military serve as proxy rulers for the economic
leaders (elites). El Salvador and Guatemala ﬁt into this pattern.

The autonomous institutional state has a highly developed infrastructure of party
or government. In Mexico, the leaders’ ideologies are required Objective environmental
factors. Thus, leaders Of institutions should be Open to suggestions regarding efﬁciency
in order to seek ways to enhance institutional viability. The political pendulum swung
from Aleman (1946-1952) on the right, Ruiz Cortines (1952-1958) on the right, Luis
Echeverria (1970-1976) on the leﬁ, Lopez Portillo (1976-1982) on the right, de 18
Madrid Hurtado (1982-1988) on the leﬁ, and so on.

Considering the frequent military regimes in this area, it is natural to see that
leaders with unforceful entry are more likely to stay in Ofﬁce for a short period. Thus,
shortly after the coups, they are withdrawn by other coup leaders, or they serve as

transitional roles and transfer the power to civilian leaders. In addition, POPDEN has a

129

negative relationship with political survival. While some variables shown above at least
explain political survival, for the strength Of survival the only one of the freedom index
is important, although it (PARTFREE) is signiﬁcant at the .10 level.

Finally, the leaders in Oceania are tested, and none of the variables are shown
to be signiﬁcant in either the selection equation or the substantive equation. The
sample size of Oceania is not large enough to use a maximum likelihood estimation
and, thus, the analysis on a model in this region is not appropriate. In the typical ML
estimation procedure, a large sample size is needed because the desirable properties of
the MLE are justiﬁed only in large sample situations (Eliason, 1991, p.8)."5

In regional comparison with the selection equations in the interaction model
(Table 10), a log likelihood ratio (-2LLR) is used as a criterion for overall signiﬁcance
of the equations, while the percentage of correctly predicted and ROE are available for
a test for the goodness-Of-ﬁt Of the model (See Table 8). Except Oceania, these
statistics can be applied. Since the substantive equation includes just 27 cases and the
selection equation is composed with 41 cases, its analyses are excluded. In Asia,
Africa, and Latin America, the selection equations are signiﬁcant at the .01 level,
whereas the selection equation for the Middle East is signiﬁcant at the .05 level and
Europe and North America at the .10 level.

Among all regions, the Middle Eastern model represents considerably the best
goodness—Of-ﬁt (ROE = 60.46) and 83.3% correctly predicted. On the contrary, the

Latin American model shows the lowest goodness-of-ﬁt (ROE = 14.47). However,

 

" It is not speciﬁed how large a sample size is enough for MLE. Considering degrees of freedom
and a level, a sample size of more than 60 is usually large enough. If d.f. = 60 and a = .05, student-t
and the standard unit normal are 2.00 and 1.96 respectively.

130

Table 10 Goodness-Of-Fit of Regional Models

Political Survival (Selection Equations) :

 

 

Region N Modal for % Correctly ROE -2LLR
NOT-SHORT (%) Predicted

Asia 120 62.50 76.67 37.79 66.04""

Africa 145 67.50 73.79 19.35 65.98"“

Middle East 102 57.84 83.33 60.46 53.44"

Europe/N. America 229 56.33 67.69 26.01 4256"

Latin America 224 66.07 70.98 14.47 82.56'"

 

" signiﬁcant at .10
” signiﬁcant at .05
"W signiﬁcant at .01

Strength of Political Survival (Substantive equations) :

 

 

Region N % Correctly -2LLR
Predicted

Asia 75 65.33 50.68"

Africa 97 53.61 50.10"

Middle East 56 83.93 77.99""

Europe/N. America 129 66.67 59.02"

Latin America 148 87.16 162.70‘“I

 

" signiﬁcant at .05
"'" signiﬁcant at .01

131
among the substantive regional models, the Latin American model (87.16%) and the

Middle Eastern model (83.93%) have a better goodness-of-ﬁt than the other models
according to the percentage of correctly predicted. However, only one variable
(PARTFREE) is signiﬁcant in the substantive equation of the Latin American mOdel. In
the case of the African model, the substantive equation represents several signiﬁcant
exogenous variables, although the model itself shows the lowest percentage of correctly

predicted (53.61%).

Conclusion

The reduced models represented the effect of each variable in three subgroups
(social backgrounds, career backgrounds, and institutional attributes). In all three
reduced models, the exogenous variables, which had signiﬁcant effect on political
survival, lost their inﬂuence on the strength of political survival. Prior to the
interactional effect, each reduced model shows a different aspect of goodness of ﬁt.

While the social background model and the institutional attributes model
improved their goodness-of-ﬁts, the career background model reduced its percentage of
correctly predicted. In other words, the career background model explained best for
political survival (SHORT or NOT-SHORT), whereas the institutional model accounted for
the strength of political survival better than others. As a result, the previous political
careers and political activities are very important to the leaders when they are elected as
top executives, but maintaining ofﬁce longer depends on institutional attributes rather
than their personal qualities.

In the interaction model, we could see the different results in comparison with

132
those in the reduced models. Although the interaction model showed better goodness-of-

ﬁt than the reduced models, the relationships of the personal backgrounds and
institutional attributes to political survival and the strength of political survival did not
appear as much clearly as in the reduced models. Rather, the results presented one
interesting ﬁnding. The characteristics of political system, i.e. , party system and type
of government, are more important for the degree of survival rather than situational
condition such as freedom status and economic freedom.

The regional models provided some different interpretations the we could not find
in the two-equation interaction model. In the African and the Middle Eastern models,
similar independent variables have signiﬁcant inﬂuence on both political survival and the
strength of political survival due to the cultural similarities and regional closeness.
However, the goodness-of—ﬁt measures indicated that the Middle Eastern model predicted
leaders’ duration much better than the African Model. It implicated that Middle Eastern
countries are more stable than African countries.

In the Latin American model, more independent variables had impact on political
survival rather than the strength of political survival. Those variables are deﬁnitely
related to the number of coup d’etat in that area. Meanwhile, in the case of European
and North American model, its goodness-of—ﬁt is lower than other regions since most
countries in that areas are liberal democratic. A slight increase of prediction in the

strength of political survival stemmed from the inclusion of the communist countries.

CHAPTER VI

Descriptive Analysis of Personal Backgrounds

In general, most descriptive statistical analyses employ graphical, tabulated, and
numerical techniques for summarizing the information in a collection of data. Sample
description is an essential task. We can not really explain or predict something unless
we can describe it clearly and accurately. In this study, even though a few of the
personal background variables are ordinal, most of them are nominal and the strength
of the association is explained by Goodman and Kruskal ’s tau, lamda or Cramer ’s V’,
etc. In addition, in the case of the tables with ordinal variables, other statistics such as
gamma, Kendall ’s tau, or Somer ’s (1, etc. are available for the relationships.

However, the main reason to employ descriptive (analyses here is to compare
personal characteristics based on institutional and systematic differences, rather than to
examine whether those variables have signiﬁcant relationships or how strong they are.
For this reason, especially in this study, calculating homogeneity scores as well as using
graphical comparisons are more helpful to cut through the superﬁcial idiosyncratic
nature of the group to uncover the personal characteristics of leaders by typologizing
leaders on the theoretically important dimensions such as region, type of head of
government, duration of ofﬁce, and number of entry.

Here, the descriptive analyses focus on the elements that contribute to political
success and give the details of speciﬁc factors. This endeavor yields discussions about

what kinds of social backgrounds political leaders have and what the similarities and

133

134

differences are. By using homogeneity scores for the personal background variables
such as education, occupation, father’s occupation, age, etc., chief executives can be
compared across different regions or different institutions. Many scholars assert that
political leaders differ from the rest of the population of the world by their social
backgrounds, career backgrounds, and personality and skills.

However, it is argued that political leaders themselves have several similarities.
Based on group comparisons rather than individual comparisons, we can test the
hypothesis that chief executives are everywhere signiﬁcantly homogeneous. Calculating
mean scores of homogeneity in each group of personal backgrounds, it is possible to
compare similarities and differences on the basis of different dimensions.

There are four-dimension comparisons in homogeneity scores, accompanied by
graphical analyses. They include categories of duration, region, type of the head of
government, and the number of entry. Also, in each dimension, it is feasible to
compare three different personal backgrounds: social backgrounds, career backgrounds,
and skills and personality. In fact, even though personality and skills data should not
be included in a quantitative empirical method, that is, a previously discussed two-
equation model, these variables can be analyzed without considering validity issues in
this part for a descriptive perspective or for the readers’ interest. Considering the logic
of homogeneity scores, even though the percentage of missing data for certain variables

is very high (see note 13), comparisons are possible without losing those variables.

135

A. Comparison by Region

Regional comparisons reﬂect cultural differences. There exists an attitudinal link
between leadership and followership that depends on national character and political
culture. Followership in different countries is more or less determined by prevailing
norms concerning the exercise of political leadership. Culture is deﬁned as the set of
properties we attribute to the system by observing the distribution of psychological
attributes among the individuals who comprise the system (Singer, 1968, p.141). For
example, even though North America is located in a different region from Europe, it
is included in the European cultural background. North America was at one time a
British colony. Canada has retained its connection with Britain and also certain aspects
of the British constitution, while the US. constitution has reﬂected the political climate
in 18th century France.

The homogeneity scores of leaders’ age in Asia and Europe presented in Table
11 are higher than other regions since more leaders from those areas became heads of
government in their 505 and 605. However, in Aﬁica and Oceania, younger (less than
50s) leaders had more chances to become heads of government. Gerontopolitics effect
is applied better to the former areas because leaders of new countries are somewhat
younger than leaders of older countries. Generally speaking, middle age (403 and 50s)
is the most popular age for being a head of government in all regions (see Figure 4).

Latin American leaders have the highest score on place of birth because a higher
proportion of leaders come from urban areas rather than rural areas. More interestingly,
while leaders in Asia (57.9%), Africa (58.9%), and Oceania (58.1%), are more likely

to come from rural areas, those in EurOpe (57.3%) and Latin America (69.0%) come

136

Table 11 : Homogeneity Scores in Social Backgrounds, Career Backgrounds, and Skills
and Personality By Region

 

Homogeneity Scores for:

 

 

 

 

 

Characteristics Asia Africa Middle East Europe Latin Amer. Oceania
Social Background :

Age .101 (145) .041 (149) .041 (111) .131 (247) .083 (292) .051 (45)
Place of Birth .032 (140) .079 (148) .009 (107) .032 (238) .148 (242) .050 (43)
SES .130 (115) .043 ( 80) .020 ( 75) .025 (181) .072 (165) .052 (20)
Father’s Occupation .083 (108) .095 ( 76) .105 ( 72) .129 (172) .156 (136) .104 (21)
Education .651 (144) .289 (150) .675 (111) .458 (246) .794 (292) .354 (44)
Major .123 (138) .066 (130) .094 (104) .214 (215) .120 (279) .203 (35)
Mean Score .195 .110 .171 .173 .255 .146
Career Backgrounds :

Foreign Experience .147 (144) .397 (150) .131 (111) .030 (246) .044 (293) .225 (44)
Military Experience .424 (142) .085 (150) .097 (111) .889 (245) .113 (296) .834 (45)
Prison/Exile .094 (142) .117 (150) .108 (111) .255 (245) .266 (296) .683 (45)
Political Activity .071 (142) .211 (150) .059 (111) .062 (245) .486 (296) .757 (45)
Previous Occupation .082 (144) .225 (150) .148 (111) .083 (245) .144 (294) .122 (45)
Experience in Congress .232 (145) .023 (149) .016 (111) .800 (245) .013 (281) .548 (45)
Experience in Ministry .470 (145) .394 (149) .511 (111) .349 (243) .226 (288) .531 (43)
Party Membership .451 (145) .142 (149) .202 (111) .952 (245) .249 (282) .548 (45)
Entry Mode .317 (145) .009 (150) .097 (111) .921 (247) .148 (297) .766 (47)
Mean Score .255 .178 .152 .483 .188 .559
Skills and Personality :

Oral Skills .159 (128) .413 (133) .291 (81) .129 (201) .213 (199) .231 (33)
Organization Skills .023 (128) .153 (133) .031 ( 81) .042 (201) .154 (199) .032 (33) -
Language Skills .435 (128) .802 (133) .291 ( 81) .597 (201) .533 (199) 1.000 (33)
Party ID .120 (145) .126 (150) .168 (111) .140 (247) .112 (295) .199 (47)
Parents’ Death/Separate .795 (128) .699 (133) .688 ( 81) .537 (201) .548 (199) .587 (33)
Physical Deformity .395 (128) .514 (133) .404 ( 81) .354 (201) .449 (199) .501 (33)
Power Connection .520 (128) .472 (133) .648 ( 81) .942 (202) .640 (199) .886 (33)
Mean Score .345 .448 .350 .373 .361 .474
Total Mean Score .267 .246 .212 .373 .252 .435

 

137

Figure 4 Age Groups by Region [ () is total number of leaders in each region.]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Asia (145)

Africa (149)

lllddle East . H (111)
Latin America (292)

N. America/Europe (247)
Oceania (45)

 

 

 

 

 

.0-39 40-49 50-59 Deot

 

 

Figure 5 Educational Level by Region

 

 

Asia (144)

 

 

 

 

 

Africa (150)

 

 

 

 

 

c111)

 

 

Middle East {“E~*

 

 

 

Latin America (292)

 

 

 

 

 

C 246)

 

 

 

N. America/Europe .

 

 

 

Oceania (44)

 

 

 

 

 

 

High and Below military Bachelor's 7A naster's E] PHD

 

 

 

 

138

from urban areas. In the case of the leader’s SES, Asian leaders are more
homogeneous than leaders in other regions. European leaders are more likely to come
from lower classes, whereas Asian and Latin American leaders come from upper
classes. Looking at father’s occupations, we can better explain the previous trend.
Even though fathers of the leaders in all regions used to be in liberal professions or
own business, especially in Asia and Latin America, they worked as party ofﬁcials,
politicians, liberal professions, owners of big farm and ranch, or businessmen, whereas
those in Europe and Oceania are more in the area of labor works and farming.

In the case of the educational backgrounds (Figure 5), leaders in Asia, the
Middle East, and Latin America, are more likely to be highly educated and represent
higher homogeneous scores. Furthermore, they possess a higher percentage of military
degrees. Meanwhile, in Africa, even though there are higher percentage of high school
and below degrees (23.3%), there also exist a higher proportion of military degrees
(29.3%). The low homOgeneity scores in Europe (16.3%) and Oceania (18.2%)
indicate that there are higher percentage of leaders with high school degree and below.

While leaders in all regions had majored in law, political science, or public
administration, at college or university, there are higher percentages of such leaders
especially in Europe (55.8%) and Oceania (51.4%) (Figure 6). Exceptionally, there are
more leaders who went to military schools in the Middle East (33.7%) and the high
percentage of leaders with military school degrees are found in Asia (18.8%), Africa
(33.1%), and Latin America (31.5%).

Career backgrounds can also provide some explanation of cultural differences.

Among the regions that have relatively higher homogeneity scores on career variables,

139

Africa has more leaders whose occupations were civil servants, militaries, liberal
professions, or owners of big business or ranches. Conversely, leaders in Europe and
Oceania are more likely to be in liberal professions or big business rather than in the
political arena. The reason for the lowest score for Europe is that the percentage of
lawyers and laborers or farmers is much higher than for other regions. Overall, as a
unit of occupation, lawyers represent as a stepping-stone for being a top executive.

Leaders in Europe (97.7%) and Oceania (93.6%) became top executives through
constitutional ways, whereas leaders in Africa (40.1%), the Middle East (35.1%), Latin
America (31.0%) and Asia (22.1%) have lower homogeneity scores since higher
percentage of leaders acquired power by forceful manners. Simply, it reﬂects that
European countries are politically more stable and developed, and Oceanic countries are
politically inﬂuenced by the European culture.

In explaining foreign experience, professional military experience, prison/exile,
and political activity, higher homogeneity scores mean that the percentage of ’Yes’
category is lower. A smaller proportion of European and Latin American leaders had
foreign experience than leaders in other regions because it results from whether the
countries were colonized before (Figure 7). These European and Latin American
countries had been independent, but those in other regions were under colonization by
some major powerful countries.

The regions that had been under colonization or wars, such as Asia, Africa, and
the Middle East, also include a higher percentage of political leaders who had
participated in political activities and experienced in prison or exile. Nonetheless, the

reason why political leaders in Europe (38.0%) were involved in political activities

140

Figure 6 Majors by Region

 

 

 

Africa E 22.3"

5.:»,.,.-,, ”3:. .I.:..w.v». ' -
2

 

 

 

 

 

 

 

 

 

 

 

 

Latin America El; 39.4 (279)
5 O
N. America/Europe 17' 55:8 ' 1 15.31 12.015)
3.7
Oceania E 51.4 2’2- (35)
Lav/PLS/PA Hed/EIIGR/Itat Sci Erin/Soc SCI/Liter
Hilttary E] Business/Econ

 

 

 

Figure 7 Foreign, Military, Prison/Exile, and P01. Activity Experiences by Region

 

 

 

Asia

Africa

Hiddle East

Latin America

N America/Europa

 

Oceania

 

 

100
(‘3

I Yoraiqn Experience Professional unitary

[:3 prison/Exile [:1 politicai Activity

 

 

 

141

did not result from the independence movements but from the underground activities
or resistance movements, such as those against Nazism or Communism. As easily as
can be predicted, Europe and Oceania prefer civilian leaders, but there are very high
percentages of professional military leaders in Africa (36.0%), the Middle East (35.1%),
and Latin America (33.4%).

Unlike the previous career experiences, in the case of political experiences, the
higher the homogeneity scores, the more ’Yes’ categories. Congressional experience
shows a comparatively much higher proportion in Europe (.800) and Oceania (.548)
than other regions. However, even though there exist differences between homogeneity
scores in ministerial experiences, generally most regions indicate almost the same
tendencies (Figure 8). The score for Latin America is a little bit lower than others since
there were more professional military men who became heads of government by force.
If we consider the inclusion of Defense Minister, Chief of Staff, or Chief of
Commander into ministry position, there might be more sudden‘coups in Latin America.

Europe (98.8%) and Oceania (86.7%) have a higher percentage of leaders who
have party experience due to systematic recruitment of political leaders, that is,
parliamentary system, prime minister system, or general secretary system. However,
leaders in Africa (68.5%), the Middle East (72.1%), and Latin America (74.8%) do not
have as much in the way of party membership as the other regions above. This is
because they have been young political systems and, thus, have not established strong
party systems. In addition, professional military leaders did not need to be in a party
in order to reach to a top executive position.

Theoretically, the primary resources for developing political institutions in any

142

Figure 8 Political Experiences by Region

 

 

 

Asia

 

Africa

 

Middle East

 

Latin America

 

N America/Europe

 

Oceania

 

 

 

 

 

 

 

 

Afrlca

Hiddle East

Latin America

N. America/Europe

Oceania

       

      

  

my

, MWMWWWW

m
w,

 

    

32
(Hm: . a. 40 3

 

 

- Ora] suns

 

Organization 51: a! I:

Llnguaqn Skll ls

 

 

 

143

country is the political skills of its politicians. The political skills needed for
developing a viable and self-sustaining political system involve the ideological
commitment, the capacity to respond to a new challenge, and the arts of administration,
negotiation, and representation. The lower the homogeneity scores for skills and
personality (Table 11), the more leaders are included in ’Yes’ categories. However, it
is shown that theory does not always match with reality.

In Figure 9, relatively more leaders in Europe (32.3%) and Asia (30.5%) have
oral skills than those in Africa (18.0%), the Middle East (23.5%), Oceania (27.3%), and
Latin America (27.1%). In the case of organizational skills, Aﬁ'ica and Latin American
leaders have higher scores (i.e. more leaders without organizational skills) because they
have more military leaders. Exceptionally, many leaders in the Middle East have
organizational skills (43.2%) and language skills (23.5%).

While the leaders in Asia and Africa have relatively more physical deformity and
family/power connection, European and Latin American leaders had been more
inﬂuenced by the fact that their parents died while they were young or that they were
separated from their parents early in life. On the contrary, family or power connection
is pretty inﬂuential in Asia, Africa, and Latin America, but not at all in Europe and the
Middle East. Although, in most regions, leaders manifest their party identiﬁcation, in
Africa, party identiﬁcation is not as important as much as other regions.

By and large, regional comparison shows several interesting ﬁndings. First of
811, Latin American leaders are the most homogeneous in their social backgrounds,
rendered by place of birth (urban), father’s occupation (politicians, landowners) and

education (above college including military). Secondly, in the case of career

144

background, Oceanic and European leaders have relatively higher homogeneous scores
than those in other regions. It means that countries in both regions have mostly
parliamentary systems and, therefore, the peaceful exit mode is more common. While
leaders in these regions do not have foreign or military experience, they have
experience in political areas, particularly, congress and party.

In the aspect of skills and personality, Oceanic and African leaders usually do
not have major skills and some sources for establishing their leadership personality. In
the case of Africa, a main reason for the higher homogeneous score is that all three
skills indicate the lowest percentages compared to other regions. In addition, although
the proportions of oral skills and organizational skills are higher, language skills are the
main contributor to a higher homogeneous score. In sum, Oceanic leaders are the most
homogeneous leaders in their social, career backgrounds, and personality and skills.
Since low score represents that the leaders have certain skills, high homogeneous score
should be interpreted that there are no skillful leaders. European leaders rank the
second in terms of homogeneity which indicates that leaders from parliamentary

political systems come from similar social and career backgrounds.

145

B. Comparisons by Duration

Based on the highest homogeneous score in age in Table 12, the leaders who
stayed very long (21 +) are more likely to become heads of government during their
405. This can be explained by the age effect (see ﬁgure 10). If someone starts his
leader’s position at younger age, he is more likely to stay in ofﬁce longer. In the case
of birth place, while shorter leaders came from urban area, leaders who came from rural
area tended to stay in ofﬁce longer. In a same fashion, SES, education, and father’s
occupation can be explained easily. The longest leaders seem to be from lower classes
because their fathers’ occupations were mostly those of laborers or farmers (44.4%).
As a result, these leaders could not get higher educations. Among other social
background variables, educational degree shows an interesting point in that the longest
leaders are likely to have less than high school degrees (36.4%) even though college
degrees almost seem to be a requirement.

In explaining career backgrounds, all outcomes seem to be easily predicted by
common sense. The reason why leaders with the duration of 3 to 10 years got into
ofﬁce through constitutional ways depends on the limited tenure of political
requirement. Some countries allow the leaders to stay just one term (4 - 6 yrs), while
other countries allow them two sequential terms (8 - 12 years) (see APPENDIX B).
Many of the shortest (0 - 2 yrs) leaders acquire power by force or coup d’etat, and such
an entry must be regarded as illegitimate (Figure 11). As a result, those leaders just
stay for a while to transfer power to another regime or be replaced by another coup.
To give a man a lifetime in which to accumulate power is dangerous. A lifetime tenure

might even be enough to establish a dynasty - sickness, senility, degeneration, insanity

146

 

 

 

 

 

 

Table 12 : Homogeneity Scores in Social Backgrounds, Career Backgrounds, and

Skills and Personality By Duration

Homogeneity Scores for:

Characteristics 0-2 3-5 6-10 11-20 ' 21+
Social Background :
Age .064 (394) .080 (257) .053 (124) .086 (97) .185 (22)
Place of Birth .037 (346) .022 (242) .013 (120) .123 (72) .077 (22)
SES .043 (206) .044 (176) .012 ( 98) .057 (60) .152 (20)
Father’s Occupation .093 (186) .097 (156) .107 ( 93) .064 (56) .220 (18)
Education .652 (394) .648 (255) .445 (125) .349 (72) .116 (22)
Major .106 (360) .092 (242) .109 (110) .086 (61) .121 (19)
Mean Score .190 .178 .130 .131 .143
Career Backgrounds :
Foreign Experience .010 (393) .050 (257) .148 (125) .289 (72) .I72 (22)
Military Experience .288 (396) .391 (255) .229 (125) .462 (72) .432 (22)
Prison/Exile .252 (396) .228 (255) .095 (125) .090 (72) .077 (22)
Political Activity .247 (396) .283 (255) .095 (125) .041 (72) .045 (22)
Previous Occupation .096 (396) .107 (256) .073 (123) .106 (72) .098 (22)
Experience in Congress .103 (384) .285 (256) .166 (125) .163 (72) .330 (22)
Experience in Ministry .316 (387) .431 (255) .482 (125) .234 (72) .330 (22)
Party Membership .287 (386) .585 (254) .377 (123) .489 (72) .834 (22)
Entry Mode .178 (400) .436 (258) .346 (125) .234 (72) .172 (22)
Mean Score .197 .310 .223 .234 .258
Skills and Personality :
Oral Skills .263 (261) .261 (203) .199 (114) .064 (67) .243 (22)
Organization Skills .165 (261) .025 (203) .108 (114) .020 (67) .045 (22)
Language Skills .692 (261) .513 (203) .547 (114) .349 (67) .329 (22)
Party ID .045 (398) .153 (258) .153 (125) .187 (72) .311 (22)
Parents’ Death/Separate .618 (261) .646 (203) .627 (114) .679 (67) .243 (22)
Physical Deformity .392 (261) .471 (203) .341 (114) .284 (67) .432 (22)
Power Connection .719 (262) .646 (203) .654 (114) .542 (67) .684 (22)
Mean Score .388 .379 .373 .302 .327
Total Mean Score .246 .296 .246 .227 .258

 

147

Figure 10 Age Groups by Duration

 

 

 

(394)

 

 

 

 

(257)

 

 

 

 

 

( 124)

 

 

 

 

 

 

 

C72)

 

 

 

 

 

 

(22)

 

 

 

 

 

.0-39 ”40-4950-591360.

 

Figure 11 Enter Mode by Duration

 

 

 

 

 

 

 

 

 

 

 

 

 

U — C400)
3 _ (258)
6 - (125)
M - 023
29] + 31 8 C22)

 

 

 

A; Constitutional [:1 Non—constitutional

 

 

 

 

148

have tenured good leaders into evil ones; yet an executive appointed for life might not
yield power gladly and might have too much power to be dispossessed without violence
(Lijphart, 1992, p.60). In reality, even though most countries put a constitutional
limitation on tenure of head of government, it is not prone to be effective.

Civil servants and professional militaries, or liberal professions and business are
the most common professions found among the leaders regardless of duration.
However, a higher percentage (27.3%) of laborers and farmers is shown in the longest
survivors. Also, leaders who have foreign experiences stayed longer. Since the
proportion of non-experiences on professional military, prison/exile, and political
activity is higher than the proportion of experiences in each duration category, the lower
homogeneity score means more experienced (Figure 12). Consequently, the categories
of longer durations (lower homogeneity scores) indicate that there are relatively more
leaders who had experiences in prison/exile, and political activities than leaders with
shorter durations. In short, graphical presentation shows the leaders with foreign, I
prison/exile, and pOlitical activist experiences except military experiences have
tendencies to survive longer. .

Granting that experiences in congress, ministry, and party are prerequisite for a
top executive position, they are needed to be examined in detail. Whether the leaders
are short-survivors or long-survivors, remarkably higher proportion of leaders in each
duration category had been in political areas prior to their top positions. Nonetheless,
there exist some differences to be mentioned. As shown in Figure 13, the leaders who
stayed short (less than 2 years) have less experiences in congress (65.9%). In the case

of experiences in ministry, the leaders who stayed for medium range of periods (3 - 5

149

Figure 12 Foreign, Military, Prison/Exile, Pol. Activity Experiences by Duration

 

 

 

 

 

100
(I)

 

Professional Hilitary

I Foreign Experience

1:] Prison/E1119 E] Political Activity

 

 

Figure 13 Political Experiences by Duration

 

 

 

        
 

t a

 

  

ﬁl11'3‘725E‘fl':

 

 

100
(‘3

I Party Congress ninistry

 

 

 

 

150

years, and 6 - 10 years) are more likely to have ministerial experiences.

Among the political experience variables, only the party membership indicates
a trend of increasing with longevity. The homogeneity scores of the experience in party
are usually higher than any other variables. However, the longest leaders (21 +) have
the highest homogeneity scores on party membership. 21 out of 22 leaders in that
duration have party membership. It is because most leaders in this duration are general
secretaries in communist countries.

As for skills and personality (Figure 14 and Table 12), with the exception of
party ideology, a lower homogeneity score means that there is a higher proportion of
’Yes’ in skills and personality variables. The leaders (38.8%) who stayed for 11
through 20 years tend to have more oral skills than those in other categories. And, in
the case of organizational skills and language skills, there are more talented people
among the leaders who stayed longer. Distinctively, among the longest survived
leaders, there is a higher proportion (27.3%) of leaders whose parents died or separated
when the leaders themselves were young. The sources that shaped leaders’ personality
have focused on their undernourished relationship with their parents, and an emotional
emptiness and thirst for power stemmed from that relationship. For the explanation of
party identiﬁcation, the leaders with conservative party identiﬁcation are likely to stay
longer. In comparison to the leaders in other categories, the leaders with short duration
(0 - 2 yrs), are more likely to be non-partisan, independent, or central for party
ideology. This is because most of them are military leaders who have never joined
parties.

Besides social and career backgrounds as well as skills and personality variables,

151

Figure 14 Political Skills by Duration

 

 

 

     
 

s — «0 (trimaran
m

 

(x)
I Oral Skills Organization Skills Language Skills

 

Figure 15 Party System by Duration

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 _ 74 S jC4DO)

3 - 81.4 1(2583

5 _ 68.8 (125)
11 (72)
27 3 C22)

 

 

 

No Party a Single Party :1 Multiparty

 

 

 

152

some country variables are related with leader’s longevity. Usually, 1/4 or 1/5 of the
countries without parties are mostly traditional systems dominated by absolute or near-
absolute monarchies, and military regimes set up against parties. However, in this
study, they are mostly military regimes, since monarchies were excluded from the
sample. Leaders in multiparty systems are likely to stay for shorter durations, while
those in single-party systems tend to stay longer (Figure 15). Within single party
system, the political career and political roles are very risky because there is ’no place
to go’ outside the party. Only through the party can a political career be achieved, and
dismissal from the party is oblivion, purgatory, or worse. High political risk and high
rewards attract the ambitious and compel their compliance. The functions of
nominating and electing candidates for political ofﬁce are attributed exclusively to
political parties. The degree of competition and monopoly in party systems affects
political turnover and the circulation of leadership.

Especially, most countries with single-party systems are the Communist regimes
and some Third World countries (i.e. Ivory Coast, Kenya, Guinea, Zambia, Tanzania,
Syria, Iraq, Egypt, etc). Therefore, in socialist economic system, leaders tend to stay
longer. Even though capitalist economic systems are popular in other categories of
duration, the shortest duration is likely to have the highest percentage of capitalist
economic system. As a result, if a society has more freedom in politics or economy,
leaders in that society tend to stay shorter. It is interesting to ﬁnd an extraordinary case
such that the shortest leaders are found in high level of political right index (40%) as
well as low level (36.8%).

In comparison of homogeneity scores on personal variables by duration, the

153

shortest leaders (0 - 2 years) represent the highest Similarity on social backgrounds,
whereas leaders in 3-5 years are relatively similar in career backgrounds. However,
except for party ideology, the longer-survived leaders have lower homogeneity scores
in skills and personality because they possess more skills and personality segments than
others. Total mean scores ﬁnally indicate there is not much difference between the

degrees of similarity of the leaders.

154
C. Comparisons by Type of Head of Government

Among the 997 top executives since 1945, there are only 12 females and 60%
of them (7 leaders) are prime ministers and the others (5 leaders) are presidents. ‘7
None of the female leaders were ever general secretaries. Even though they were
elected by national assembly or national election, some of them had ﬁlled the place of
their late husbands or some were offsprings of well-known former prime ministers.
However, more than one-half of them had been elected through a competition with male
candidates. Surprisingly, it is rare to ﬁnd female head of governments in advanced
liberal democracies, but rather they are mostly from developing or underdeveloped
countries. Exceptionally, until now, there are no female leaders found in African
countries.

According to the homogeneity scores in Table 13, especially with respect to the
social background variables, presidents and prime ministers have similar characteristics,
but general secretaries require different explanations. General secretaries are more
likely to be older when they arrive at the top (Figure 16). Even though the scores of
birth place are same for all types of leaders, general secretaries (56.3%) are from rural
areas and others (55.2% for Presidents and 52.2% for prime ministers) from urban
areas.

While most presidents (40.4%) and prime ministers (44.7%) came from upper

 

’7 Female top executives are Isabella Peron (1974 - 1976) in Argentina, Lidia Guelier Tejada

(1979) in Bolivia as an interim leader, mary Eugenia Charles (1980 - ) in Dominica, Indira Ghandi (1967 -

1977, 1980 - 1984) in India, Golda Meir (1969 - 1974) in Israel, Agatha Barbara (1982 - 1989) in Malta,

Violeta Chamorro (1990 - ) in Nicaragua, Gro Harlem Bnmdtland (1981, 1986 - 1989) in Norway, Benezir

Bhutto (1988 - 1991) in Pakistan, Corazon Aquino (1987 - 1992) in Philippines, Sirimaro Bandarike (1960
- 1965, 1970 - 1977) in Sri Lanka, and Margaret Thatcher (1979 - 1990) in England.

155

classes, general secretaries (61.9%) had lower class backgrounds since their fathers’
occupations were mostly laborers or farmers (65.8%). They could not get as much
support for education by their fathers as other types of leaders. Consequently, high
school degrees are more common among general secretaries (35.4%), whereas the others
have more college degrees. There is a high percentage of militarily educated presidents
(33.1%). Because military leaders are included in presidents, there are more
nonconstitutional exits than other types. Most leaders, regardless of types of leadership,
study political science, law, or public administration. However, interestingly, medicine
or engineering (34.3%) are found just as much as aforementioned among general
secretaries.

Among the different occupations previously mentioned, liberal professions or

business owners are commonly found in all types of head of government. Actually, civil

Figure 16 Age Groups by Type of Head of Government

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

President 20-8 (562)

Prime llinister 29 B (379)

General Secretary 31 3 (48)
We-asII4o-4950-59D50+

 

 

 

 

156

Table 13 : Homogeneity Scores in Social Backgrounds, Career Backgrounds, and Skills
and Personality By Type of Government and Number of Entry

 

Homogeneity Scores for:

 

 

 

 

 

Characteristics President Prime General Once 2 +
Minister Secretary
Social Background :
Age .038 (562) .077 (379) .131 (48) - .045 (861) .165 (128)
Place of Birth .013 (506) .005 (364) .036 (48) .004 (793) .036 (125)
SES .014 (340) .041 (254) .210 (42) .010 (542) .090 ( 94)
Father’s Occupation .093 (303) .080 (244) .375 (38) .082 (499) .096 ( 86)
Education .630 (562) .548 (377) .104 (48) .561 (860) .587 (127)
Major .094 (524) .185 (342) .124 (35) .085 (786) .150 (115)
Mean Score .166 .169 .156 .147 .199
Career Backgrounds :
Foreign Experience .085 (563) .020 (377) .082 (48) .050 (860) .086 (128)
Military Experience .077 (563) .810 (378) .771 (48) .258 (861) .613 (128)
Prison/Exile .153 (563) .347 (378) .023 (48) .222 (861) .078 (128)
Political Activity .258 (563) .180 (378) .063 (48) .209 (861) .105 (128)
Previous Occupation .140 (563) .089 (378) .102 (48) .134 (861) .108 (128)
Experience in Congress .004 (553) .689 (376) .695 (47) .111 (849) .532 (127)
Experience in Ministry .351 (557) .487 (376) .024 (46) .315 (851) .663 (128)
Party Membership .173 (553) .818 (376) .844 (48) .344 (850) .821 (127)
Entry Mode .068 (568) .755 (381) .572 (48) .225 (869) .638 (128)
Mean Score .145 .466 .354 .207 .404
Skills and Personality :
Oral Skills .263 (421) .149 (309) .235 (45) .241 (670) .075 (105)
Organization Skills .138 (421) .021 (309) .027 (45) .063 (670) .075 (105)
Language Skills .499 (421) .630 (309) .683 (45) .610 (670) .574 (105)
Party ID .101 (566) .140 (381) .947 (48) .100 (867) .118 (128)
Parents’ Death/Separate .634 (421) .620 (309) .548 (45) .641 (670) .515 (105)
Physical Deformity .479 (421) .361 (309) .324 (45) .444 (670) .262 (105).
Power Connection .604 (421) .737 (310) .834 (45) .670 (671) .659 (105)
Mean Score .375 .372 .538 .384 .319
Total Mean Score .216 .364 .361 .243 .326

 

157

servants or military professions (42.3%) are the most popular for the presidents,
whereas lawyers (23.3%) are second most for the prime ministers and laborers or
farmers (27.1%) for the general secretaries. Regardless of the types of leaders, there
are more leaders with foreign experience than those with nonexperience. Experience
in prison/exile (52.1%) seems to be common to the general secretaries and it results
from frequent political activities (60.4%).

In comparison of political experience, working in congress seems to be very
important to the leaders who want to be prime ministers or general secretaries.
Meanwhile, although being a congressman seems to be a springboard to political

success, working as a Minister is more important in presidential systems and prime

Figure 17 Political Skills by Type of Head of Government

 

 

Pre51dent V

Prime Minister

General Secretary

 

 

 

I Oral sxiiis Organization sniis Language Skills

 

 

 

158

minister systems. Of course, party membership is very important in all types of
leadership, but in a presidential system its signiﬁcance is reduced because military coup
leaders easily become top executives.

Comparisons of skills and personality do not provide any signiﬁcant differences
of homogeneity scores. Nevertheless, more prime ministers (30.0%) have oral skills,
general secretaries (53.3%) have organizational skills, and presidents (14.3%) have
rather language skills (Figure 17). In terms of the personality variables, general
secretaries have more than other types, especially physical deformity and separation or
death of parents. However, presidents had more power/family connection than other
types. Party identiﬁcation does not explain much at all since general secretaries are
from the systems mainly controlled by communist parties.

Overall comparisons indicate that overall social backgrounds do not provide any
serious impacts on leaders’ similarities, that prime ministers and general secretaries are
more similar in their career backgrounds, and that general secretaries have relatively
higher homogeneous scores on skills and personality than other types of heads. In
general, presidents are more heterogeneous than prime ministers or general secretaries.
With respect to the electoral procedure of top executives, people may choose more
diverse leaders by direct popular elections or national assemblies, whereas indirect

electoral systems generally prefer top executives with similar backgrounds.

159
D. Comparisons by Number of Entry

In this part, the homogeneity scores of Table 13 provide the comparisons
between the leaders who became heads of government only once and those who did so
more than twice. Number of entry does not give as much in terms of the difference as
other factors that are compared previously. However, it is necessary to check out some
variables that contribute some inﬂuence on their similarities.

In case of the social background variables, there does not exist much difference
between the two groups of leaders. The leaders who became top executives more than
twice have higher scores on age, since they started during their 408 for the ﬁrst time
and thus for subsequent terms their ages are getting older than the ﬁrst time. There are
more leaders from upper class (50.5%) among the leaders who have been in office more

than twice.

Figure 18 Foreign, Mil., Prison/Exile, and P01. Act. Exp. by Number of Entry

 

 

Once

 

 

more Than Twice

 

 

 

 

 

0 20 40 60 90
(*3

'1r Pr012531ona1 Hilitary

 

I Foreign Experience

Prison/Exile Political Activity

 

 

 

160

Figure 19 Political Experiences by Number of Entry

 

 

 

Once

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\
more Than Twice 874
5
g l r l r l r l r
U 20 41] 60 80 100
(“J
I Party Congress ﬂinistry
Figure 20 Personality by Number of Entry
Once
More Than Twice
3O
(90

I PhySical Delorlanca Parents' Death/Separation

Power Connection

 

 

 

161

On the contrary, it is easy to ﬁnd signiﬁcant differences with career
backgrounds. The reason why the leaders who have been in ofﬁce more than twice have
higher scores on constitutional exit mode is that previously experienced leaders might
have gotten out of ofﬁce in peaceful ways and thus got into ofﬁce for the second time
or more via constitutional ways. Quite a few of one-term leaders have gotten into
ofﬁce through coups or forceful ways. It purports other trends: the scores on military
experience is low for the leaders who have been in ofﬁce once, since many of them got
there by forceful ways.

In the opposite to military experience, prison/exile experience or political activity
seems to be a big advantage to be in ofﬁce again (Figure 18). Also, leaders with more
than two terms have more experience in congress, ministries, and parties (Figure 19).
As a result, to be a head of government more than twice is related to political
experiences or political activities rather than military experiences.

While the leaders who became top executives more than twice have more oral
skills than the leaders who did once, the other skills (organizational skills and language
skills) do not show much difference. However, the leaders with more than twice have
a relatively higher proportion of all three personality characteristics, i.e., parents’ death
or separation, physical deformity, and power/family connection (Figure 20).

Here again, the similarities in career as well as social backgrounds are easily
found among the leaders who became top executives more than twice. Nonetheless,
even though the leaders who were in ofﬁce only once possess higher similarities, those
similarities are toward the direction of non-possession of skills and personality. In

other words, as we assumed, the leaders with more than two terms comparatively have

162

more skills and the causes of personality formation. Above all, the homogeneity scores
of the leaders based on number of entry tell us that leaders with more than two terms

are more similar to each other than one-term leaders.

CHAPTER VII

CONCLUSION

The decline of authoritarian rule in Southern Europe, Latin America, and other
regions in the 19708 and 1980s, as well as the collapse of communism in Eastern
Europe and Soviet Union in the beginning of 1990s, may represent the most signiﬁcant
and encouraging historical trends of the turbulent and often tragic twentieth century.
Authoritarianism is now on the wane and democracy on the rise across the globe in
spite of widespread mid-century pessimism about the prospects for democratic
governance (Rustow, 1990). As the expectation for the rise of civilian and
constitutional rule grows, leadership change may be an important prospect in political
science.

As mentioned by Huntington (1991, p.315), the two key factors affecting the
future stability and expansion of democracy are economic development and political
leadership. Democracy contributes to stability by providing regular opportunities for
changing political leaders and changing public policies. An institutional or tactical
analysis of political development is usually related to a crucial problem for leaders -
how to survive in power. As long as politics is deﬁned as a process that is deeply and
intricately interwoven between governing and power struggle, political leadership may
be explained rationally based on the stability and security of the system and irrationally
on the basis of motivation for getting into top positions in politics. Thus, the chief
executive is the lead actor in the political drama of contemporary society, since the

position refers to the political head of government rather than to the ceremonial chief

163

164

of state.

According to Confucius’ thought and the pre-Christian philosopher’s
perspectives, "If the ruler is virtuous, the people will also be virtuous." Even though
classical literature has provided several qualities and conditions for ideal leaders, the
image of political leaders has changed over the years. The qualities of leaders
deﬁnitely are no longer important for their successful survival. As long as the selection
procedures are democratic, people choose their leaders by participating in elections. In
addition, political culture is better conceived not purely as the legacy of the communal
part but as a geological structure with sedimentary deposits from many historical ages
and events (Diamond, 1993, p.412). In each new historical period, new value
orientations have partially displaced but not completely erased preexisting ones.

As one of the dimensions on political performance, durability as a concept of
stability should be considered. Especially in political leadership, that is, a position of
a head of government, its stability deﬁnitely relates to its longevity. If political stability
is deﬁned as stability and continuity of basic political institutions and processes
themselves, institutional longevity which measures the long-term stability of political
structures should be based on the leaders’ longevity. The length and terms of tenure
must be sufﬁcient to permit the holders to perform their ofﬁces, but not so long as to
permit tenure to become entrenchment in ofﬁce. They must be secure enough to permit
incumbents to govern, but not so secure as to encourage complacency or corruption in
Ofﬁce.

The study of political stability does not require justiﬁcation in terms of the

importance of the phenomena related to it. Rather it requires some establishment of

165

boundaries within which the process of interest can be studied. Sometimes it will be
shown that the most stable systems, whether domestic or international, are at the same
time the most predictable ones in a very simple sense - the ability based on performance
to predict the performance of that system at the same future time. Political Success,
actually, depends on performances in ofﬁce as well as "rally-around-the-ﬂag effects"
from international crises. Being the right person is almost as important as "being in the
right place at the right time" (Sirnonton, 1984, p.165).

However, since failures of the system are fault of incompetent power-holder, the
situation can be greatly improved by changing the incumbent authorities. The resulting
stability in the administrative structure makes it possible to have the everyday lives of
ordinary citizens to remain largely unaffected even by the changes of national leaders.

Therefore, the power of the people to control government is the negative power to
remove a government from ofﬁce and replace it. As Simon mentioned (1965),
"Political leadership is like a bus driver whose passengers will leave him unless he
takes them in the direction in which they wish to go, they leave him only minor
discretion as to the road to be followed." As a result, nondemocratic structure may
frequently be relatively stable. In case of the democratic structure, it is deﬁnitely true
that being a great power is no longer fun (Schoenbaurn, 1973).

The effort to base legitimacy on performance sometimes gives rise to what can
be termed the performance dilemma. Indeed, predictions on duration in office can be
different from legal duration even where leaders are elected for a ﬁxed term, especially
in constitutional presidencies. For example, in many Latin American countries,

presidents often have to resign or are dismissed by coups before the end of their legal

166

term. In addition, it is possible to see more changes of top executives in countries
whose political institutions are less stable.

Meanwhile, the length of a particular presidency may depend partly on personal
inﬂuence, namely on the president’s ability to get himself renominated and reelected.
No American president in nearly half a century, not even Eisenhower, managed to
sustain such high level of public trust. Reagan was the ﬁrst president since Eisenhower
to ﬁnish eight full years in ofﬁce and the ﬁrst one since Andrew Jackson to leave ofﬁce
safe in the knowledge that his vice president had been elected to succeed him
(Etheredge, 1993, p.149). Sometimes, the transfer of political power never resulted
from a set of accepted, well-deﬁned, institutionalized procedures. For example, in
China prior to 1949, the individuals in collective leadership - and hence the serious
contenders for Mao’s mantle of power - had to possess some particular qualiﬁcations.

The main answers for this study in empirical aspects would be focused on two
possibilities: (1) Is it possible to employ personal characteristics as well as situational
characteristics to construct a composite proﬁle of individuals who, in a given society,
are likely to be successful leaders (SHORT or NOT-SHORT) ?; (2) Is it possible to utilize
them to arrive at a composite proﬁle of individuals who turn to be the different level
of successful leaders (Who are more successful? or How longer do leaders stay? ) ?
In addition, through descriptive analyses, similarities and differences on personal
characteristics of the leaders, including skills and personalities, have been discussed on
the basis of region, duration, type of head of government, and number of entry.

In this study, 17 hypotheses have been tested using a two-equation model in an

effort to see the relationships between personal backgrounds and institutional

167

characteristics and political survival and its strength as political success. Three reduced
models - social backgrOund model, career background model, and institutional model -
were examined to see the effect of independent variables in each model and to compare
which model explains and predicts better for political survival. An interaction model,
which includes all explanatory variables from all three reduced models, has been tested
for ﬁnding which makes leaders stay in ofﬁce for a short period or not (selection
equation) and for how long (substantive equation). Even though 17 hypotheses are
mainly tested, actually each model has more tests since several dummies are included
because most independent variables are nominal.

All models have evaluated what elements are appropriate for explaining and
predicting duration of ofﬁce as the indicators for political survival of heads of
government across the countries since 1945 until 1992. Predictability is conﬁned to the
large aggregates of data over long time periods. Whereas all selection equations contain
984 leaders after data manipulation, only 581 leaders are included in all. substantive
equations after the selection procedure. In addition, regional models have been tested
to see cultural effects.

In each reduced model, some independent variables proved statistically
signiﬁcant or possibly signiﬁcant only in the selection equation, that is political
survival. However, none of them proved signiﬁcant in the substantive equation for the
strength of political survival. Meanwhile, in the interaction model, more independent
variables turned out to be signiﬁcant in the substantive equation than in the selection
equation. All individual effects on political survival in each reduced model are almost

similar to those in the interaction model. So, in this conclusion, the discussion will be

168

focus on the individual effects from interaction model.

One of the most interesting ﬁndings concern the vanishing effect of signiﬁcant
independent variables when explaining the strength of political survival in all three
reduced models. The reason why some independent variables explain only political
survival and not the strength of political survival is due to honeymoon effects that all
political leaders face. New leaders do plan to affect public policy during their
honeymoons because they have the desire and the capacity to do so during their terms.
Therefore, after a certain short period (0 - 2 years), known as the honeymoon period,
social and career backgrounds of the leaders deﬁnitely tend to be less important as
much as in election times or the selection procedures. Those personal backgrounds
seemed to be crucial for the people to elect or select the leaders on the basis of the
leaders’ image established by their personal characteristics, not based on their political
performance, that is, institutional characteristics.

In general, in explaining political survival (selection equation) itself, at least one
of the dummy ,variables for each of the independent variables on social and career
backgrounds, except entry mode (NONCONSTI) proved statistically signiﬁcant for the
hypothesis testing. Conversely, most institutional characteristics, such as country age
(OLD), party system (SINGLEP and MULTIP), homogeneity index (HOMOGEN), population
density (POPDEN) and region did not prove signiﬁcant at all. Those variables which
were found to explain best why leaders stay short or not, are mostly personal
backgrounds of the leaders before entering their top executive ofﬁce.

All three social backgrounds, such as age, education, and birth place were

statistically signiﬁcant, although directions of the relationships were opposite to the

169

expected for college degree and urban born. As one would expect, at least one of the
dummies of career backgrounds, i.e., political experience (POLALL), foreign experience
(FORSTUD and FORBOTH), and political ideology (CONSERV) were proven to be
signiﬁcant, but previous occupation (LAWPOL) had a negative relationship with political
survival. Those who had been lawyers and politicians before entering into a top
executive position are probably more vulnerable to withdrawal by sudden military
coups. Moreover, those professions are more favorable in liberal democracies with
prime minister systems in which leadership transfers are very frequent.

Type of government (PREZ) as an indicator of institutional difference and literacy
rate (LITERACY), freedom status (FULLFREE), and economic freedom (ECONFREE) as the
indicators of modernization proved statistically signiﬁcant or possibly signiﬁcant.
Surprisingly, economic freedom is yet negatively related to political survival. Perhaps,
as an indicator of modernization, it tells instead that leaders in the societies with low
economic freedom are more likely to stay in office longer since those societies are not
economically stable, and, thus, leaders tend to remain in power until they boost the
economy up to a certain stage that makes the polity stable both economically and
politically.

Nonetheless, in comparison with other two reduced models (selection equations),
the career background model shows the best goodness-of-ﬁt. In other words, in order
to explain and predict whether leaders stay in ofﬁce for a short period (0 - 2 Yrs.) or
not, the previous experiences in foreign countries and political arenas as well as
political ideology, that is, orientations toward politics, are the most important factors.

Even though each three reduced model proved Signiﬁcant for the test of the whole

170
model using -2LLR in both selection model and substantive model, the hypotheses

testing on each independent variable are not available in the substantive equation of
each reduced model.

Consequently, in order to ﬁnd the relationships with a number of the indicators
above and the strength of political survival (substantive equation), the results from the
interaction model are inspected. While social background variables did not show any
difference from the selection equation, some in career backgrounds and some in
institutional attributes provide somewhat different outcomes. When comparing the
ﬁndings of the selection equation with those from the substantive equation, the
following structure of results could be observed.

First of all, all dummies concerning political experience in career backgrounds
did not prove statistically signiﬁcant at all. To the leaders who already have spent two
more years in ofﬁce, political experience is irrelevant to the prediction of the strength
of political survival. Instead, their political performance during their tenure is probably
much more vital. Second, country age and party systems became important in
explaining the strength of political survival. The ﬁndings indicate that the leaders in
young countries and in single party systems are more likely to maintain power longer.
By and large, the young countries have not been in a ﬁrm base of political stability
since their independence from colonization, and usually a founder of the national party
or a ﬁrst leader of the country would have maintained his office until the new country

was settled down as a stable polity.I

 

2' For examples, they are Pindling in Bahamas (1967 - ), Kharna in Botswana (1966 - 1980),
Ahidjo in Cameroon (1960 - 1982), Pereira in Cape Verde (1975 - ), Gouled Aptidon in Djibouti (1977 -
), Jawara in Gambia (1965 - ), Sekou Toure in Guinea (1958 - 1984), Houphouet-Boigny in Ivory Coast

171

Third, freedom status and economic freedom indicating situational differences
have lost their signiﬁcance and thus could not predict the strength of political survival
any more. The leaders who already have stayed in ofﬁce over 2 years might be
exempted from the threat of political and economic instability. Rather, more
institutional factors, as mentioned earlier, such as country age, party systems, and type
of head of government, are very supportive in predicting the strength of political
survival.

Above all, the interaction model showed the best goodness-of-ﬁt among the
models that have been discussed in both the selection equation and the substantive
equation. Additionally, it is more possible to test the relationships between independent
variables and political survival and the strength of political survival, which the
substantive equation in each reduced model by itself could not ﬁnd. However, based
on the success table (Table 7) as well as the percentage of correctly predicted, the
model itself might be more desirable to predict political survival than to predict the
strength of political survival.

Although regional dummies have never been signiﬁcant in the models, dividing
the interaction model by region has shown different ﬁndings. African and Latin
American models include mainly more personal variables which are signiﬁcant when
explaining political survival, whereas for the strength of political survival Middle
Eastern and African models showed several signiﬁcant indicators not only for personal

background but also for institutional attributes. Overall, while the Middle Eastern

 

(1960 - ), Kenyatta in Kenya (1964 - 1978), Banda in Malawi (1964 - ), Senghor in Senegal (1960 -
1981), Pinto da Costa in Sao Tome and Principe (1975 - ), Nyerere in Tanzania (1961 - 1985), Lini in
Vanuatu (1980 - ), Kaunda in Zambia (1964 - ), etc.

172
model has the best goodness-of-ﬁt in explaining political survival itself, the Latin

American and Middle Eastern models are proven to be the best in accounting for the
strength of political survival.

The question, whether the data from diverse cultures and time periods are
functional, should be overcome by minimizing as much as possible the introduction of
ethnocentric or otherwise biased materials. However, justiﬁcation of gathering social
information for the comparative study of political leadership seem to withstand most
criticisms on empirical inefﬁciency. In contrast to the general elite studies, it is much
easier and more possible to collect reliable social background and career paths of the
leaders as heads of government since they are much better known and publicized than
other elites groups. The only one problem is that it is very difﬁcult to collect and
operationalize psychological information. While it is possible to ﬁnd leaders’ personal
histories, past achievements, and basic political philosophies, their skills or personalities,
such as hobbies, habits, and weak points, do not receive much attention, and, therefore,
it is not easy to ﬁnd some crucial psychological factors. Hence, it is not easy to
identify the individual traits that contribute to leader’s success.

Many scholars believe that the study of political leadership as we ﬁnd it in the
context of contemporary social psychology is no longer a study of the mighty, of the
personalities and behaviors of the outstanding ﬁgures of modern society, but of the
many minor heads, bosses, and superiors of everyday, mainly institutional, life. If
’ﬁtness’ for a certain position becomes less evolutionary and more social, reputation or
status - family and class background, education, profession, money, etc. - are substitutes

for real competitive proofs. By ’worth’ is meant that set of qualities and conditions that

173

society values in assigning social class rank; it is what the electorate admires. For
example, those that in the past have rated highest are; in China, scholarly learning and
prowess at killing games; in India, Brahman birth; in Europe, royal birth (Girvin, 1989).
However, instead of parents’ worth, personal worth, that is recognized socially as
personal attributes, is included in the new formula for worth.

Some background characteristics have more relevance than others for political
success within same political systems. A priori use of background characteristics as
explanatory variables has different effects for different types of heads of government
in different regions of the world. However, gross similarities between the social
backgrounds or career backgrounds of heads of government can not be used as proof
of political success or failure and the degree of their success. Signiﬁcance of those
backgrounds for political success is greater in developing societies than in societies with
more stable political institutions. Therefore, the impact of social and career
backgrounds on political success remains plausible, but ambiguous and unsubstantiated.

It would seem that long leadership will tend to prevail in countries where the
extent of social and economic change is low. Elsewhere, the occurrence of long
leadership would hinge on the presence of special conditions countering the propensity
of leaders to be rapidly weakened by social change. Even though biographies often
endeavor to give descriptive accounts of the personality characteristics of the political
leaders, it is premature to begin discussing the impact of personality as we did not
know which ’traits’ were relevant to political success.

Like any quantitative researches, in this study, we can ﬁnd that measures of

duration of ofﬁce are limited to the year unit and typically insensitive to some validity

174

problems. By nature, the sample, here, is a censored sample in which some
observations on the dependent variable corresponding to known sets of independent
variables are not observable. Using the duration of ofﬁce as the dependent variable
poses new problems. First, the value of the dependent variable is unknown or censored
for the leaders who ere not out of ofﬁce during the observed period. The second
problem is how to include explanatory variables that change in value over the
observation period. There is no satisfactory way of incorporating time-varying
explanatory variables in a multiple regression predicting time of an event.

As a solution for the former problem, in order to reduce the problem of missing
data on duration ofﬁce resulted from censoring, the leaders who are still in ofﬁce during
the observed period need to be given elaborate consideration. They make up 11.5 %
(115 cases) of the total leaders examined (997 cases) and excluding them from the
analysis is not a reasonable solution. Thus, considering the theoretical aspects, only the
short durations (0 - 2 yrs) during the observed period are excluded. It is not a perfect
solution for missing data, but at least it prevents from losing many available
information.

Most situational explanatory variables incur the second problem stated above
since their value are changing over time. If continuous variables themselves, such as
freedom status, economic freedom, literacy, population density, age of country, etc., are
considered as annual data, the measurements of the year when the leader became a head
of government are used. However, the unavailable data are replaced with the values
of the closest year even though they are not accurate. However, in case of the

stationary variable, i.e., homogeneity index, it is reasonable to use given values of a

175

certain year when the leader started his ofﬁce as a head of government since its change
over the years is not seriously enormous. _

Although we investigate as many potential causes of successful political
leadership as the available data allowed, some important variables - SES, father’s
occupation, major in school, or skills and personalities - have undoubtedly been
omitted from the causal system. If something related to them could be found, it might
be general enough across time and space to lend itself to predictive testing when applied
to contemporary and future leaders, men for whom extensive biographical data would
not be available, but for at least some of whom sufﬁcient articulated data might be.
Thus, these data are just analyzed in a descriptive way.

If the results of the estimation provide justiﬁcation or corroboration for
theoretical perspectives on political survival, then it could be argued that this model has
led to some progress in the ﬁeld. Not only disproof, invalidation, lack of progress, but
also satiation due to overresearching an area are the most frequent motives for the
discontinuous nature of historical change in the sciences. Many people assert that the
greatest weakness of research in this ﬁeld has not been a lack of facts but the disparity
between theory and empirical research. Despite masses of accumulated data, it is
necessary to ﬁnd rather few general relationships between demographic, career
background, and situational factors and political survival as political success. Andreas
Pickel (1989, p.90) asserts, "We should try to improve and expand our knowledge by
criticizing rather than by justifying our theories, by searching for refutations rather than
conﬁrmations." Even though some stand up more successfully to criticism than others,

and this is why they can be considered closer to the truth than others, constructive

176

search for refutation still forces us to develop new and better theories.

The study of political survival as political success based on leader’s duration in
ofﬁce is important in at least a couple of reasons. First, individual leaders have
different qualities associated with being top executives which people expect, despite
social backgrounds or career backgrounds. Second, polities differ in the economic and
political output of the government, such as economic freedom, political system, freedom
index, party system, type of head of government, and so on. Therefore, as Rousseau
mentioned, it is true that the strongest is never strong enough to always be the master
unless he transforms strength into right and obedience into duty.

Finally, there is some motivation that individual leaders need to stay longer as
top executives. However, these psychological aspects can not be generalized by some
collected biographical information because of the lack of validity and reliability of the
sources. Personal attributes of leaders have a large effect. They relate to deeper
elements of the personality or to events that occurred in childhood. Leaders have to be
able to grasp what is important and what is trivial, to weigh alternatives, to foresee
possible outcomes. Decision-making a matter of will power and courage, the leader has

to have the emotional ability to stop the debate in his mind and take a particular side.

RECOMMENDATION

If it is possible to ﬁnd other information on certain background variables, such
as birth order, sibling combination, religion, and ethnicity, etc., it will provide an even
more interesting explanation on political success. According to Stewart (1977), the

relationship between birth order and political Zeitgeist is explained as follows: F irst-

177

born may have superior odds of success in times of crisis, with the proviso that middle
children be more disposed toward political leadership (They are more likely to lead in
periods of calm.) and later-born children toward revolution. These tendencies may arise
from childhood patterns of sibling relationship (Galton, 1874: Rejai and Phillips, 1983).

As Weber repeatedly emphasized, it is not necessarily what the leader is but how
he is perceived by his followers that is decisive for the validity of charisma (Willner,
1968, p.4). In so far as charisma can be seen as a quality of an individual, it lies in his
capacity to project successfully an image of himself as an extraordinary leader. In other
words, charisma is a certain quality of an individual personality. In biographies or the
documents about the political leaders, it may not be difﬁcult to ﬁnd the description
indicating charismatic personality - acuteness, eye for detail, ideological ’founding
father’ or ’initially’ charismatic leader, discriminatory powers of the observer and
recorder, nonverbal behavior. With this subject, the research should be narrowed down

to the certain limited target population.

APPENDICES

APPENDIX A

Data Collections and Data Codebook

A. Data Sources

The most important data sources for collecting information on the leaders as
heads of government in each country and their duration of ofﬁce are Heads of State and
Government (1985), Elections since 1945: A Worldwide Reference Compendium (1989),
World Atlas of Elections 1986 (1987), The International Almanac of Electoral History
(1991), etc.

The main sources of social background variables are Current Biography, New
York Times Biographical Edition that accumulates the articles of Man of the News,
Keesing '3 Contemporary Archives, Current World Leaders (1960 - ), Almanac of
Current World Leaders, Encyclopedia of World Biography, Who ’s Who in the World,
International Who 's Who, The Annual Obituaries (1980 - ), The Britannica Yearbook,
Who 's Who in Modern World: 1860 - 1980, Historical Dictionary series of each
country, and several biographies from some speciﬁc regions or countries, such as
Biographical Dictionary of Latin American And Caribbean Political Leaders,
Dictionary of Irish Biography, Encyclopedia of Latin America, Who ’s Who in Africa,
Makers ofModern Africa, Encyclopedia of Asian History, Who ’ Who in Australasia and
Far East, etc.

The sources for institutional data are World Quality of Life Indicators, World
Development Report, Britannica World Data, Political Parties of the World, and World

Handbook of Political and Social Indicators III: 1948 - 1982, World Urbanization:

178

I950 - 1970, Freedom in the World, 1974 - 1990, The New Book of World Rankings,

and etc.

179

180

B. Operationalization and Measurement

Dependent Variable

Duration of Office (DURATION or DURACAT) : For a two-equation model, duration
of ofﬁce as a continuous variable is divided into 4 categories (DURACAT).
And, for the descriptive analysis, it is categorized into 5 values (DURATION).

DURATION: DURACAT:
O=O-2 O=0-2
l=3-6 l=3-5
2=7-12 2=6-10
3=13+ 3=11~20
4=21+

Indemndent Variables

(a) Social Backgrounds of Individual Leaders

Age (when the person became a leader)(AGE or AGECAT) : Age is used as a
continuous variable in the two-equation model, whereas it is categorized for the
descriptive analysis.

O=Under40
l=40-49
2=50-59
3=60and0ver

Place of Birth (BPLACE):

O = Rural + Foreign Country

1 = Urban (Cosmopolitan areas) + Capitol
Father’s SES (SES):

0 = Low

1 = Middle
2 = Upper

181

Degree of Education (DEGREE): It is operationalized in two different ways:
DEGREE (l) for the descriptive analysis and DEGREE (2) for the two-
equation model.

DEGREE (l) DEGREE (2)
0 = Below High School 0 = Below High School
1 = Coll/Univ. (BA, BS, LLB) 1 = Above College

2 = Master’s (MA, MS)
3 = Doctorate (LLD, MD, PhD)
4 = Military

Major of Education (MAJOR):

O = Law/PLS/Diplomacy/Public Administration

1 = Medicine/Pharmacy/Engineering/Nattual Science/Agriculture
2 = Journalism/Language/Education/Social Science/Theology

3 = Business/Accounting/Finance/Economics

4 = Military

Father’s Occupation (FATHOCCU): Since it is very difﬁcult to collect the
information of father’s occupation, ’Not Available’ is included as one of the
categories of the variable.

0 = Lawyer

1 = Civil Servant/Military

2 = Party ofﬁcial/Politician/Political Activist/Union Activist

3 = Journalist/Teacher/Professor/Liberal Profession/
Business/Own Farms fTechnical/White Collar/Medical Doctor/
Engineer/Accountant

4 = Manual Laborer/ Small Wage Earner/Farmer/Agriculture

5 = Not Available

(b) Career Backgrounds
Party Membership (PARTEXP): If a leader had joined in a party, it is coded as 1.

0=No
1=Yes

182

Experience of Ministerial Position or a Chief of Staff (MINIEXP): If someone has
been in ofﬁce as a minister of a certain department or in almost the same level
of a powerful position such as a Chief of Staff in military or a Commander of
army, it is coded as l. I

0=No
1=Yes

Experience in Congress (CONGEXP): Regardless of the differences between lower,
and upper house or senators and congressmen, if someone has been elected for
Congress, it is coded as 1.

0=No
1=Yes

Political Experience (POLEXP):

0 = None

1 = Party or Congress or Ministry Only

2 = Party + Congress/Party + Ministry/Ministry + Congress
3 = Party + Congress + Ministry

Career before getting into Politics (CAREER):
CAREER (1) for the two-equation model

0 = Lawyer

1 = Civil Servant/Military

2 = Party Ofﬁcial/Politician/Political Activist/Union Activist
4 = Manual Laborer/ Small Wage Earner/Farmer/Agriculture

CAREER (2) for the descriptive analysis

0 = Lawyer/Politician/Political Activist

1 = Civil Servant

2 = Military

3 = Journalist/Teacher/Professor/Liberal Profession/
Business/Own Farms fTechnical/White Collar/Medical Doctor/
Engineer/Accountant

4 = Manual Laborer/ Small Wage Earner/Farmer/Agriculture

183

Experience in Prison or Exile (PRISEXIL): If a leader had experience in prison or
in exile for political reasons, and not for personal reasons, it is coded as l.

0=No
1=Yes

Military Service (MILEXP): Even though someone had been in obligatory military
service, he is included in ’no experience’ category. The leader who was a
professional military, is coded as l.

O = No
1 = Professional Military

Political Activity (POLACT): It decides whether someone participated in underground
movements, political activity at schools, or other activities related to politics.

0=No
1=Yes

Entry Mode (ENTRY): Constitutional way of entry includes the entry by the national
referendum, by a council or committee, or an elected chamber or majority vote
in parliament, whereas non-constitutional way comprises self-appointment

(Usurpation) or military coup.

O = Constitutional
1 = Non-Constitutional

Foreign Experience (FOREIGN): Measurement rule depends on whether someone has
been abroad for studying, working, travelling, or living for a long term.
FOREIGN (l) for the descriptive analysis and FOREIGN (2) for the two-equation

model.
FOREIGN (l) FOREIGN (2)
0 = No 0 = None
1 = Yes 1 = Study in Foreign Countries

2 = Work in Foreign Countries
3 = Both

184

(c) Institutional Attributes

Region (REGION): It is based on geographical division as well as historical or
cultural similarities. As a result, even though North African countries are
located in Africa, they are included in Arabic territory. North American countries
are combined with European countries.

0 = Asia
1 = Africa
2 = Middle East

3 = Europe/North America
4 = Latin America ‘
5 = Oceania

Party System (PARTYSYS):

O = No Party
1 = One Party
2 = Multi-Party

Type of Head of Government (LEADTYPE):

LEADTYPE (l) LEADTYPE (2)
O = President/Military O = Prime Minister/Gen. Sec.
1 = Prime minister 1 = President/Military

2 = General Secretary

Economic Freedom (ECONFREE):

1 (Not Free) - 5 (Fully Free)

Population Density (POPDEN): The unit of population density as a continuous
variable is 1,000 per Sq. Km. Since it is changing year by year, the value of
population density is based on the year when a leader became a head of
government.

Homogeneity Index (HOMOGEN): It is an index that combines racial, religious,
linguistic, and ethic diversity. The data during 1960 to 1965 is used since there
is no big change in values.

185

Country Age (COAGECAT):
0 = Independent Since 1960 (0 - 30)
l = Independent Before 1960 (31 +)

Freedom Status (FREESTAT): It is an index of combination of political rights and
civil rights indices.

0 = Low
1 = Medium
2 = High

Urbanization (URBAN): It is related to the process of mobilization: we can divide the
countries into small-states, industrialized, and developed countries. It is measured
by percentage of population in cities of over 100,000 people or more. It also
changes year by year.

Literacy (LITERACY): Data for most countries relate to populations aged 15 and
over; but in some cases, the ﬁgures are based on population over age 6. It is
widely used for determining social and cultural development.

d) Personal Dispositions

Oral Skills (ORAL): If leaders are described by the words such as eloquent
speaker, master of persuasion, brilliant orator, powerful voice, or ability to
quickly grasp oral and written arguments, they have oral skills.

0=No
1=Yes

Organizational Skills (ORG): If leaders are described by the words such as disposition
to command with inﬂuencing purpose and boundless energy, or a man of true
ability, they have organizational skills.

0=No
1=Yes

186

Language Skills (LANG): If they speak more languages other than their native
language, they have language skills.

0=No
1=Yes

Political Ideology (PARTYID) : A political leader is almost always possessed of some
set of philosophic and political convictions that he has, for one reason or another,
made his own. It is usually the same as the ideology of the party that he joins.

0 = None

1 = Liberal

2 = Independent (Center)
3 = Conservative

Physical Deformity (PHYSDEF): If a leader has physiological disposition or bad
health, his personality development is inﬂuenced by it.

0=No
1=Yes

Parental Death or Separation in Early Age (FADIESE):

0=No
1=Yes

Family or Power Connection (FAMCONN):

0=No
1=Yes

(d) Exit Mode (REMOVAL): The constitutional exit includes the cases that are not
involved in forceful removal, while non-constitutional exit means forceful
removal from the ofﬁce. The former category has automatic rule (end of term),
completion of special task, defeat in election, natural death, and voluntary
resignation or retirement. The latter includes the leaders who are assassinated,
and deposed by military coup or other ways. Death, if nonviolent, may safely be
counted as ’extra—political’, but illness is often political face-saving device for an

187

involuntary replacement. If prime ministerial changes are to be taken into
account, the most straightforward criterion is to count all of them without trying
to measure the exact degree to which an illness may be political (Lijphart, 1984,
pp.268-269). Nonetheless, natural deaths or illness is counted as constitutional
for the study.

0 = Constitutional
1 = Non-constitutional

Hazard Rate

Hazard Rate

1 88
APPENDIX B

Hazard Rate (Life Table)

Figure 21 Hazard Rate of Total Sample (N=984)

 

 

Total (984)

 

 

 

 

l l l, l l l l l l I, l l L, l L, l l l l l 1 l l

 

O 2 4 6 6 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44

Duration (Years)

Figure 22 Hazard Rate of Censored Sample (N=882)

 

 

 

 

 

 

 

 

Existing Total (882)
T Censored (102)
1 1 1 1.; 1 1 L 1 l 1 1 1 1 1 1 1 1 1 L

 

 

 

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44

Duration (Years)

Table 14 Ordinary Least Squares Estimation

1 89
APPENDIX C

 

Dependent Variable: DURATION

 

 

 

Variable Coefficient Standard Error t-Statistic
Constant 10.13656 2.54477 3.98330‘"
AGE -0.1 1623 0.021 14 -5.49769"“
COLLEGE -1 .08637 0.64495 -1 .68444‘
URBAN -1 .04769 0.41331 4.53484"
LAWPOL -3.38199 1.13111 -2.98997""
CIVSERV -4.27806 1 .24275 -3.44242"'"
MILITAR - 1 .50243 1 .23204 -1.21946
LIBTECH -3.31387 1.10553 -2.99753”‘
POLONE 0.40314 1.30247 0.30952
POLTWO 0.8545 1 1.37807 0.62008
POLBOTH 0.81863 1 .3 8737 0.59006
F ORSTUD 1.23355 0.51625 2.38945”
FORWORK 0.59810 0.58392 1.02428
F ORBOTH 1.50344 0.65585 2.29236"
NONCONSTI 0.33858 0.66742 0.50729
LIBERAL 2.08950 0.90527 2.30814“
CENTRE 1.75370 1.15409 1.51955
CONSERV 2.92347 0.86769 3.36929“
OLD -1.49812 0.57361 -2.61 174'"
PREZ 1 .90623 0.56955 3.34689‘”
SINGLEP 3.27892 1.04722 3.13107""
MULTIP -0.09300 0.96138 -0.09673
POPDEN 0.00078 0.00094 0.83 106
HOMOGEN -0.00044 0.00108 -0.40554
LITERACY 0.0203 1 0.01 151 1.76524“
ECONSTAT -0.33774 0.28799 -1.17277
PARTFREE 0.66352 0.75779 0.87560
FULLFREE 1 .09754 1.01599 1 .08027
ASIA 0.99248 1.23797 0.80170
AFRICA 0.09237 1.28693 0.07178
MIDEAST -0.44012 1.25672 -0.35021
EURNAME 1 .37779 1.09892 1 .25377
LATINAME -0.36371 1.18041 -0.30812
Number of Observations 871
R-squared 0.20412
Corrected R-squared 0.17373
Sum of Squared Residuals 27714.8
Standard Error of the Regression 5.75087
Durbin-Watson Statistic 1.85944
Mean of Dependent Variable 5.25373

 

‘ signiﬁcant at .10
*" signiﬁcant at .01

" signiﬁcant at .05

LIST OF REFERENCES

LIST OF REFERENCES

Agassi, Joseph. 1986. Scientiﬁc Leadership. In Changing Conceptions of Leadership,
edited by Carl F. Graumann and Serge Moscovici. New York: Springer-Verlag.

Aldrich, John H. and Charles F. Cnudde. 1975. Probing the Bounds of Conventional
Wisdom: A Comparison of Regression, Probit, and Discriminant Analysis.
American Journal of Political Science 19: 571-608.

 

Aldrich, John H. and Forrest D. Nelson. 1984. Linear Probabilig, Logit, and Probit
Models. Beverly Hills and London: Sage Publications.

Alexander, Robert J ., ed. 1988. Biographical Dictionm of Latin American and
Caribbean Political Leaders. New York: Greenwood Press.

Almanac of Current World Leaders. 1960-1990. Los Angeles, California.

Ames, Barry. 1987. Political Survival: Politicians and Public Policy in Latin America.
Berkely: The University of California Press.

The Annual Obigm. 1980-1991. Chicago and London: St. James Press.

Baker, Randall, ed. 1992. Public Administration in Small and Island States. West
Hartford: Kumarian Press.

Ballard, Elizabeth J. and Peter Suedfeld. 1988. Performance Ratings of Canadian
Prime Ministers: Individual and Situational Factors. Political Psychology, Vol.9,
No.2: 291—302.

Beck, Carl, et a1. 1973. Comparative Communist Political Leadership. New York: D.
McKay C.

Bell, David S., Douglas Johnson, and Peter Morris, eds. 1990. Biographical Dictionm
of French Political Leaders since 1870. New York and London: Simon &
Schuter.

Bienen, Henry and NIcholas Van De Walle. 1991. Of Time and Power: Leadership

190

191
Duration in the Modern World. Stanford: Stanford University Press.

Bishop, Yvonne M., Stephen E. Fienberg, and Paul W. Holland. 1975. Discrete
Multivariate Analysis: Theory a_r;_d Practice. Cambridge, MA. and London,
England: The MIT Press.

Blondel, Jean. 1980. World Leaders: Heads of Government in the Postwar Period.
London and Beverly Hills: Sage Publications.

Blondel, Jean. 1982. The Organizational of Governments: A Comparative Analysis of
Governmental Structure. Beverly Hills: Sage Publications.

Blondel, Jean. 1984. Dual Leadership in the Contemporary World: A Step towards
Executive and Regime Stability? In Comparative Government and Politics,
edited by Dennis Kavanagh and Gillian Peele. London: Heinemann.

Blondel, Jean. 1987. Political Leadership: Towards a General Analysis. Beverly Hills:
Sage Publications.

Boylan, Henry. 1988. A Dictiongy of Irish Biography. 2nd ed. Gill and Macmillan,
Ltd.

Britannica Boon the Year: People of the Year. 1938-1992. Chicago: Encyclopedia
Britannica, Inc.

Bullitt, Stimson. 1977. To be a Politician. Revised edition. New Haven: Yale
University Press.

Bunce, Valerie. 1981. Do New Leaders make a Difference?: Executive Succession and

Public Policy under Capitalism and Socialism. Princeton, N.J.: Princeton
University Press.

Burns, James MacGregor. 1977. Wellsprings of Political Leadership. American
Political Science Review LXXI, 1 (March): 266-275.

 

Burns, James MacGregor. 1979. Leadership. New York, Hagerstown, San Francisco,
London: Harper & Row, Publishers.

Calvert, Peter. 1983. Politics, Power and Revolution: An Introduction to Comparative
Politics. Brighton: Wheatsheaf.

Calvert, Peter, ed. 1987. The Process of Political Succession. New York: St. Martin’s
Press.

Campbell, A., et a1. 1960. The American Voter. New York, N.Y.: Wiley.

192

Camp, Roderic A. 1979. Women and Political Leadership in Mexico. The Journal of
Politics Vol.4]: 417-441.

 

Camp, Roderic A. 1980. Mexico’s Leaders: Their Education and Recruitment.
Tucson, Arizona: The University of Arizona Press.

Carlyle, Thomas. 1966. On Heroes, Hero-Worship and the Heroic in Histog. Edited
by Carl Anthony Niemeyer. London: University of Nebraska Press.

Carter, April. 1979. Authorig and Democracy. London: Routledge and Kegan Paul.

Cell, Charles P. 1974. Charismatic Heads of State: The Social Context. Behavior
Science Research 9(4): 255-305.

 

Charlton, Roger. 1986. Political Leadership: Executives in the Modern World. In

Political Realities: Comparative Government, edited by Roger Charlton. London
and New York: Longrnan Ltd.

Christensen, Ronald. 1990. Log-Linear Models. New York: Springer-Verlag.

The Columbia Dictiom of Political Bioggaphy. 1991. New York: Columbia
University Press.

Constantini, Edmond and Kenneth H. Craik. 1972. Women as Politicians: The Social
Background, Personality and Political Careers of Female party Leaders.
Journal of Social Issues 28(2): 217-236.

 

Crewe, 1., ed. 1974. British Political Sociology Yearbook, Vol.1: Elites in Western
Democracy. London: Croom Helm.

Current Bioggaphy. 1940-1990. New York: H. W. Wilson Co.
Current World Leaders: Bioggaphy and News. 1960-1990. Pasadena, California.

da Graca, John V. 1985. Heads of State and Government. New York University
Press.

Dahl, Robert and Edward R Tufte. 1973. Size and Democracy. Stanford: Stanford
University Press.

Day, Alan and Henry W. Degenhardt, eds. 1980 Political Parties of th_e World.
Longrnan Group, Ltd., A Keesing’s Reference Publication.

de Blij, Harm J. 1967. Systematic Political Geoggaphy. New York: John Wiley &

193

Sons.

Daganzo, Carlos. 1979. Multinomial Probit: The Theory and Its Application to
Demand Forecasting. New York and London: Academic Press.

Delpar, Helen, ed. 1974. Encyclopedia of Latin America. The McGraw-Hill.

Diamond, Larry, ed. 1993. Political Culture and Democracy in Developing Countries.
Boulder: L. Rienner Publishers.

Dickie, John, and Alan Rake. 1973. Who’s Who in Africa: The Political Milim and

Business Leaderﬂf Africa, London: Wheatsheaf House.

Dodd, Lawrence C. 1984. The Study of Cabinet Durability: Introduction and
Commentary. Comparative Political Studies, Vol. 17, No. 2 (July): 155-161.

Dion, Léon. 1968. The Concept of Political Leadership: An Analysis. Canadian
Journal of Political Science 1: 1-17.

 

Dogan, Matti, ed. 1988. Comparing Pluralist Democracies: Strains on Legitimacy.

Boulder and london: Westview Press.

Dogan, Matti, ed. 1989. Pathways to Power: Selecting Rulers in Pluralist Democgies.
Westview Press.

Deutsch, Karl W. 1961. Social mobilization and Political Development. American
Political Science Review 55: 493-514.

 

Dhrymes, Phoebus J. 1986. Limited Dependent Variables. In Handbook of .
Econometrics Vol.3. edited by Zvi Griliches and Michael D. Intriligator.
Amsterdam: North-Holland.

 

Ebel, Robert. 1972. Governing the City-State: Notes on the Politics of the Small Latin

American Countries. Journal of Inter-American Studies and World Affairs,
14:325-346.

Eckstein, Harry. 1979. Suppprt for Regime: Theories and Tests. Research Monograph
No.44 . Center for International Studies.

Eckstein, Harry. 1992. Regarding Politics: Essays on Political Theory, Stabilig and
Change. Berkeley: University of California Press.

Edinger, Lewis J. 1964. Political Science and Political Biography: Reﬂections on the
Study of Leadership (1). The Journal of Politics Vol.26, No.2 (May): 423-439.

 

194

Edinger, Lewis J. 1964. Political Science and Political Biography (11). The Journal
of Politics Vol.26, No.3 (August): 648-696.

 

Edinger, Lewis J., ed. 1969. Theog and Method in Comparative Elite Analysis. New
York.

Edinger, Lewis J. 1974. Where are the Political Superstars? Political Science
Quarterly, Vol.89, No.2 (June): 249-268.

Edinger, Lewis J. 1975.. The Comparative Analysis of Political Leadership.
Comparative Politics, Vol.7, No.2 (January): 253-269.

Edinger, Lewis J., ed. 1976. Political Leadership in Industrial Societies: Studies of
Comparative Analysis. Huntington & New York: Robert E. Krieger Publishing

Company.

Edinger, Lewis J. 1990. Approaches to the Comparative Analysis of Political
Leadership, The Review of Politics Vol.52: 509-523.

 

Edinger, Lewis J. and Donald D. Searing. 1967. Social Background in Elite Analysis:
A Methodological Inquiry. American Political Science Review Vol.61 (June):
428-445.

 

Edinger, Lewis J. and Donald D. Searing. 1967. Leadership: An Interdisciplinary
Bibliography. In Political Leadership in Industrial Societies, edited by Lewis J.
Edinger, 348-366. New York.

Elderveld, Sammuel James. 1989. Political Elites in Modern Societies: Empirical
Research and Democratic Theog. Ann Arbor: University of Michigan Press.

Eliason, Scott R. 1993. Maximum Likelihood Estimation : Logic and Practice. A
Sage University Paper.

Embree, Ainslie T., ed. 1988. Encyclomrlia of Asian Histog, Vol.1 - Vol.4. New
York: Charles Scribner’s Sons.

Encyclomdia of World Biography. 1973. The McGraw-Hill.

Erickson, Erik B. 1962. Young Man Luther: A Study in Psychoanalysis and History.
New York: Norton.

Etheredge, Lloyd S. 1993. The Hardball Practitioner. In Political Psychology, edited
by Neil J. Kressel. New York: Paragon House Publishers.

Eulau, Heinz and Moshe Czudnowski, eds. 1976. Elite Recruitment in Democratic

195

Politics: Comparative Studies Across Nations. Beverly Hills: Sage
Publications.

Fenno, R. 1959. The President’s Cabinet. Cambridge, Massachusetts: Harvard
University Press, Vintage Edition.

F iedler, Fred E. Leadership. 1971. Pamphet and New York: General Learning Press.

Field, G. Lowell, and John Higley. 1980. Eliticism. London, Boston and Henley:
Routledge and Kegan Paul.

Finer, S. E. 1988. The Man on Horseback: The Role of the Milim in Politics. 2nd
ed. Boulder, Colorado: Westview Press.

F isichella, Domenico. 1984. The Double Ballot System as a Weapon against Anti-

System Parties. In Choosing an Electoral System: Issues and Alternatives, edited
by Arend Lijphart and Bernard Grofman. New York: Praeger.

Force, George T. and Jack R Van der Silk, eds. 1969. Theog and Research in the

Study of Political Leadership: An Annotated Bibliography. Carbondale:
Southern Illinois University Press.

 

Foti, et a1. 1982. Effects of Leadership Labels and Prototypes on Perceptions of
Political Leaders. Journal of Applied Psychology, 67 (June): 326-333.

Gastil, Raymond D. 1974-1990. Freedom in the World. New York: Freedom House.

 

George, Alexander. 1974. Assessing Presidential Character. World Politics 26
(June): 234-282. '

Gibbs, Cecil A. 1947. The Principles and Traits of Leadership. Journal of Abnormal
and Social Psychology, 42: 267-84.

Gibbs, Cecil A. 1954. Leadership. In Hgdbook of Social Psychology , Vol.2, edited
by Gardner Lindsey. 877-920. Reading, Mass.

Girvin, Brian. 1989. Change and Continuity in Liberal democratic Political Culture.
In Contemmrm Political culture: Politicg in at Postmodern Age, edited by John
R. Gibbins. London: Newburry Park.

Gorvin, Ian, ed. 1989. Elections since 1945: A Worldwide Reference Commndium.
Longrnan International Reference.

Greenstein, Fred. 1986. Personality and Politics. In Political Leadership: A Source
Book, edited by Barbara Kellerman. Pittsburgh: University of Pittsburgh

196

Press.

Grob, Leonard. 1984. Leadership: The Socratic Model. In Leadership:
Multidisciplinarv Perspectivgg, edited by Barbara Kellerman, with a foreword by
James MacGregor Burns. Englewood Cliffs: Prentice-Hall, Inc.

Gunson, Phil, Greg Chamberin, and Andrew Thompson, eds. 1991. The Dictiopary
of Contemporm Politics of Central America and the Caribbean. Simon and

Schuster.

Gupta, Diak K. 1990. The Economics of Political Violence: The Effect of Political
Instabilig: on Economic Growth. New York: Praeger.

Gurr, Ted R. 1974. Persistence and Change in Political Systems, 1800-1974.
American Political Science Review 68 (December): 1482-1504.

 

Gurr, Ted R. and Muriel McClelland. 1971. Political Performance: A Twelve-Nation
Study. Comparative Politics Series, Vol.2. Beverly Hills: A Sage Professional
Paper.

Haberman, Shelby J. 1978. Analysis of Qualitative Dar; Vol.1. New York and
London: Academic Press.

Hagle, Timothy M. and Glenn E. Mitchell 11. 1992. Goodness-of-F it Measures for
Probit and Logit. American Journal of Political Science Vol.36, No.3
(August):762-784.

 

Heckman, James J. 1976. The Common Structure of Statistical Models of Truncation,
Sample Selection and Limited Dependent Variables and a Sample Estimator for
such Models. Annals of Economic and Social Measuremenr, Vol.5, No.4: 475-
492.

Hibbs, Douglas, et a1. 1982. On the Demand for Economic Outcomes: Macroeconomic
Performance and Mass Political Support in the United States, Great Britain, and
Germany. Journal of Politics Vol.44, No.2 (May): 426-462.

 

Hecksner, Gunnar. 1960. The Study of Comparative Government and Politics.
London: George Allen & Unwin Ltd.

Hodgkinson, Christopher. 1983. The Philosophy of Leadership. Oxford: Basil
Blackwell. .

Hollander, Edwin. 1978. Leadership Dynamics: A Practical Guide to Effective
Relationship. In Leadership, edited by J. M. Burns. New York: Harper and
Row.

197

Holmes, Jack E. and Robert E. Elder, Jr. 1989. Our Best and Worst Presidents: Some

Possible Reason for Perceived Performance. Presidential Studies Quarterly,
19 (2-3): 529-588.

Hook, Sidney. 1986. The Eventful Man and the Event-making Man. In Political

Leadership: A Source Book, edited by Barbara Kellerman. Pittsburgh: University
of Pittsburgh Press.

Huntington, Samuel P. 1991. The Third Wave: Democratization in the late Twenties
Centuries. Norman and London: University of Oklahoma.

Huntington, Samuel P. 1971. The Change to Change. Comparative Politics, 111, 3
(April): 283-322.

Hurwitz, Leon. 1973. Contemporary Approaches to Political Stability: A Review
Essay. Comparative Politics, 5 (April): 449-463.

Ilchman, Warren J. and Norman T. Uphoff, eds. 1972. The Political Economy of

Development: Theoretical and Empirical Contributions. Berkeley: University of
California Press.

Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Countries. Princeton:
Princeton University Press.

International Who’s Who. 1935-1992. London: Europa Publications Ltd.

Jackman, Robert W. 1993. Power without Force: The Political Capacity of Nation-
States. Ann Arbor: The University of Michigan Press.

Jackson, Robert and Carl Rosberg. 1982. Personal Rule in Black Africa: Prince
Autocrat, Propher, Tyrant. Berkeley: University of California Press.

 

Janda, Kenneth F. 1972. Towards the Explication of the Concept of Leadership in
Terms of the Concept of Power. In Political Leadership: Readings for an
Emerging Field, edited by Glenn D. Paige. New York: The Free Press.

Janowitz, Morris. 1954. The Systematic Analysis of Political Biography. World
Politics Vol.6, No.3 (April): 405-412.

 

Jennings, Eugene E. 1960. An Anatomy of Leadership: Princes, Heroes, and
Suxrman. New York: Harper and Brothers.

Kay, Susan Ann. 1991. Introduction to the Analysis of Political Data. Englewood
Cliffs, NJ: Prentice-Hall.

198

Kellerman, Barbara, ed. 1986. Political Leadership: A Source Book. Pittsburgh:
University of Pittsburgh Press.

King, Anthony. 1975. Executives. In Handbook of Political Science Vol.5, edited by
Fred 1. Greenstein and Nelson W. Polsby. 173-256. Addison-Wesley Publishing
Company.

 

 

Kingsley, David. 1976. World Urbanization 1950-1970. Westport, Conn: Greenwood
Press. '

Kurian, George Thomas. 1984 & 1991. The New Book of World Rankings. New
York: Facts on File Publications.

Lane, Robert E. 1961. Political Life: Why People Get Involved in Politics. Glencoe,
Illinois: The Free Press.

Lao Tzu. 1961. Tao Te Ching. Translated by John C. H. Wu. and edited by Paul K.
T. Sih. New York: St. John’s University Press.

Laski, H. J. 1933. Democracy in Crisis. Chapel Hill, NC: University of North
Carolina Press.

Lasswell, Harold. 1934. Psychopathology and Politics. Chicago: University of
Chicago Press.

Lasswell, Harold. 1948. Power and Personalig. New York: Norton.

Lasswell, Harold and Daniel Lerner. 1965. World Revolutionm Elites. Cambridge,
Mass: MIT Press.

Lasswell, Harold and Abraham Kaplan. 1950. Power and Socieg: A Framework for
Political Inquiry. New haven: Yale University Press.

Lee, Sidney. 1928. Principles of Biography. In Elizabethan and Other Essays.
Oxford: Oxford University Press.

Leonard, R L. and Richard Natkiel. 1987. World Atlas of Elections 1986. London:
Hodder and Stoughton.

Lerner, Daniel. 1957. Communication Systems and Social Systems: A Statistical
Exploration in History and Policy. Behavioral Science 2 (October): 266-275.

 

Lijphart, Arend. 1977. Democrgy in Plural Societies: A Comparative Exploration.
New Haven: Yale University Press.

199

Li jphart, Arend. 1984. Measures of Cabinet Durability: A Conceptual and Empirical
Evaluation. Comparative Political Studies, Vol 17, No. 2 (July): 265-279.

Lijphart, Arend. 1992. Parliamengy versus Presidential Government. Oxford: Oxford

University Press.

Lindblom, C. E. 1968. The Policy-making Process. Englewood Cliffs, N.J.:
Prentice-Hall.

Linz, Juan J. 1988. Legitimacy of Democracy and the Socio-Economic System of

Western Democracies. In Comparing Pluralist Democracies: Strains on
Legitimacy. Edited by Mattei Dogan. pp.65-113. Boulder: Westview.

Lipschutz, Mark R. and R Kent Rasmussen, eds. 1986. Dictiom of African
Historical Bioggiphy. Second Edition. Berkeley, Los Angeles, and London:
University of California Press.

Lipset, Seymour Martin. 1981. Political Man: The Social Bases of Politics. Expanded
and Updated edition. Baltimore: Johns Hopkins University Press.

Lipset, Seymour Martin and William Schneider. 1983. The Conﬁdence Gap: Business,
Labor and Government in the Public Mind. New York: Free Press.

Little, Roderick J. A. and Donald B. Rubin. 1987. Statistical Analysis with Missing
Data. New York and Toronto: John Wiley & Sons, Inc.

Lord, Robert G. and Myron Rush. 1980. Effects of Sex and Personality on Perceptions
of Emergent Leadership, Inﬂuence, and Social Power. Journal of Applied
Pachology, 65: 176-182.

Lorvvin, Val R 1971. Segmented Pluralism. Comparative Politics, 3: 141-175.

MacDougall, Terry Edward. 1982. Political Leadership in Contemmm Japan. Ann
Arbor: Center for Japanese Studies, University of Michigan.

Mackie, Thomas T. and Richard Rose. 1991. The International Almanac 0f Electoral
History. 3rd ed. Washington, DC: Library of Congress Cataloging-in-
Publication Data.

Maddala, G. S. 1983. Limited-De ndent and itative Variables in Econometrics.
Cambridge and New York: Cambridge University Press.

McKelvey, Richard and William Zavonia. 1975. A Statistical Model for the Analysis
'of Ordinal Level Dependent Variables. Journal of Mathematical Sociology, 4:
103-120.

200

Mainwaring, Scott. 1993. Presidentialism, Multiparty System, and Democracy: The

Difﬁcult Equation. Comparative Political Studies, Vol.26, No.2 (July): 198-
230.

Makers of Modern Africa : Proﬁles in Histog. 1981. Published by Africa Journal
Limited for Africa Books Ltd.

Markus, Gregory B. and Phillip Converse. 1979. Political Environment and the
Dynamics of Public Attitudes: A Panel Study. American Journal of Political
Science 23 (May): 338-359.

 

Mayser, George and Margaret Wagstaffe, eds. 1987. Research Methods for Elite
Studies. Contemporary Social Research 14. London: Allen & Unwin.

Merriam, Charles. 1967. Four American Pm Leaders, reprinted. F reeport: Books for
Libraries.

Mishler, William and Ann Hildreth. 1984. Legislatures and Political Stability: An
Explanatory Analysis. Journal of Politics 46 (February): 25-59.

 

Montesquieu. 1977. The Spirit of Laws. With an Introduction, Notes, and
Appendixes, edited by David Wallace Corrinthers. Berkeley: University of
California Press.

Morgenthau, Hans J. and David Hein. 1983. Essays on Lincoln’s Faith and Politics,
edited by Kenneth W. Thompson. Washington, DC: University of Press of
America.

Morrison, D. G. and H. M. Stevenson. 1971. Political Instability in Independent
Black Africa: More Dimensions of Conﬂict Behavior Within nations. Journal
of Conﬂict Resolution 15: 347-368.

 

Mughan, Anthony and Samuelson C. Patterson, eds. 1991. Political Leadership in
Democratic Societies. Chicago: Nelson-Hall Publishers.

Neustadt, Richard. 1960. Presidential Power. New York: John Wiley.

I_h_e New Africans: A Guide to the Contemmrg History of Emergent Africa and Its
Leaders. 1967. London: Paul Hamlyn Ltd.

 

New York Times Biographical Edition Service. 1970-1991.

Nice, David. 1984. The Inﬂuence of War and Party System Aging on the Raking of
Presidents. Western Political Quarterly, 37 (Summer): 443-455.

201

Nietzsche, Friedrich. 1969. Thus Smke Zarathustra. Translated by R. J. Hollingdale.
Baltimore: Penguine Books.

Nordlinger, Eric A. 1972. Conﬂict Regglation in Divided Societies. Cambridge,
Mass: Center for International Affairs Harvard University.

Paige, Glenn D. 1977. The Scientiﬁc Study of Political Leadership. New York: The
Free Press.

Palmer, Alan, ed. 1980. Who’s Who in Modern World : 1860-1980. New York: Holt,
Rinehart & Winston.

Palmer, Alan, ed. 1990. The Pengu_1n' Dictiom of Twentieth-Cam Histog. Third
edition. Penguin Books.

Panebianco, Angelo. 1988. Political Parties: Organization and Power. Cambridge:
Cambridge University Press.

Perlmutter, A. 1981. Political Roles and Milim Rulers. London: Cass.

Pickel, Andreas. 1989. New Ask Who Should Rule: Karl Popper and Political Theory.
Canadian Journal of Political Science XXII, No.1 (March): 83-105.

 

Pimlott, Ben. 1990. The Future of Political Biography. The Political Quarterly,
Vol.61, No.2 (April-June): 214-224.

Pindle, David F. 1993. Head of State and Head of Government in Comparative
Perspective. Presidential Studies Quarterly, Vol.108, No.2 (Summer): 271-283.

Plato. 1968. The Republic of Plato. Translated, with Notes and Interpretative by Allan
Bloom. New York: Basic Books Inc.

Portis, E. B. 1987. Charismatic Leadership and Cultural Democracy. The Review of
Politics Vol.49, No.2: 231-250.

 

Powell, G. Bingham. 1982. Contemmm Democracy: Participatiop, Stabiliry, and
Violence. Cambridge, Massachusetts: harvard University Press.

Prewitt, Kenneth. 1970. The Recruitment pf Political Leaders: A Study of citizen-
Politicians. Indianapolis: Bobbs-Merrill.

Pridham, Georffrey, ed. 1990. Coalitional Behavior in Theorv apdr Practice: An
Inductive Mpdel for Western Eurom. Cambridge: Cambridge University Press.

Putnam, K. J. 1964. Charisma and Political Leadership. Political Studies 12: 341-

 

202
354.

Putnam, Robert D. 1971. Studying Elite Political Culture: The Case of Ideology.
American Political Science Review Vol.65 (September): 651-681.

 

Putnam, Robert D. 1973. The Beliefs of Politicians. New Haven: Yale University
Press. -

Pye, Lucian. 1985. Asian Power and Politics: The Cultural Dimensions of Authority.
Cambridge, Mass: The Belknap Press of Harvard University Press.

Rae, Douglas W. and Michael Taylor. 1979. The Analysis of Political Cleavages.
New Haven and London: Yale University Press.

Rake, Alan. 1992. Who’s Who in Africa: Leaders for the 19903. Netuchen, N.J.: The
Scarecrow Press, Inc.

Rejai, Mostafa and Kay Philips. 1988. Becoming Political Leaders. In Loyalists and
Revolutionaries : Political Leaders Compared. New York : Praeger.

Reich, Bernard, ed. 1990. Political Leaders of the Contemmra_ry Middle East and
North Africa: A Biographical Dictiom. New York: Greenwood.

The Revolution in France The Works of Edmund Burke, Vol.2. 1970. London: Henry
Bohn.

 

Riggs, Fred. 1981. The Rise and Fall of Political Development. In The Handbook of
Political Behavior Vol.4, edited by Samuel L. Long. New York: Plenum Press.

 

Rogowski, R. 1974. Rational Legitimacy. Princeton: University of Princeton Press.

Rose, Richard. 1987. Ministers and Ministers: A Functional Analysis. Oxford:
Claredon Press.

Rose, Richard. 1993. Evaluating Presidents. In Researching the Presidency: Vital
, uestions New A roaches edited by George C. Edwards 111, et a1. Pittsburgh
and London: University of Pittsburgh Press.

Rose, Richard and Ezra N. Suleiman, eds. Presidents and Prime Ministers.
Washington, DC: American Enterprise Institute for Public Policy Research.

Rosenbach, William E. and Robert L. Taylor. 1992. Milim Leadership: In Pursuit
of Excellence. Second edition. Boulder: Westview Press.

Rustow, Dankwart A. 1966. The Study of Elites: Who’s Who, When, and How.

203
World Politics Vol. XVIII, No.4 (July): 690-717.

 

Rustow, Dankwart A. 1967. A World of Nations. Washington: Brookings Institutions.

Rustow, Dankwart A. 1990. Democracy: A Global Revolution? Foreigp Affairs, 69
(Fall): 75-91.

Sanders, David and Valentine Herman. 1977. The Stability and Survival of
Government in Western Democracies. Acta Politig 12: 346-377.

Schlesinger, Joseph A. 1966. Ambition and Politics: Political Careers in the United
States. Chicago: Rand McNally.

Schumpeter, Joseph. 1975. Capitalism, Socialism, and Democracy. New York:
Harper Colophon.

Seligman, Lester G. 1950. The Study of Political Leadership. American Political
Science Review Vol.44, 4 (December): 904-915.

 

Seligman, Lester G. 1964. Leadership in New Nation. New York: Atherton Press.

Singer, J. David. 1973. Man and World Politics: The Psychological Interface. Journal
of Social Issues. 24 (3) (July): 127-156.
California: Clio Press Ltd.

Sheehan, T. 1976. Senior Esteem as a Factor of Socio-Economic Complexity.
Gerontology 16: 433-40.

Shils, Edward. 1982. The Political Class in the Age of Mass Society: Collectivistic
Liberalism and Social Democracy. In Does Who Governs Matter? : Elite
Circulation in Contemmrgy Societies, edited by Moshe M. Czudnowski, 13-34.
Northern Illinois University Press.

Shugart, Matthew Soberg and John M. Carey. 1992. Presidents and Assemblies:
Constitutional And Electoral Dynamics. Cambridge: Cambridge University
Press, 1992.

Shumacher, Rose, eta1., eds. 1989. World Qu_a_li§y of Life Indicators. Santa Barbara,

Siedman, R B. 1978. The Web of Group Afﬁliations and Conﬂict. New York: The
Free Press.

Simon, Herbert. 1965. Administrative Behavior. New York: Free Press.

Sirnonton, Dean Keith. 1981. Presidential Greatness and Performance: Can We Predict

204
Leadership in the White House? Journal of Personalig, 49 (Summer): 306-323.

Sirnonton, Dean Keith. 1986. Disposional Attributes of Presidential Leadership.
Journal of Exp_erimental Social Psychology, 22 (Summer): 399-418.

Sirnonton, Dean Keith. 1986. Presidential Personality. Journal of Personalig and
Social Psychology, 51 (July): 149-160.

Sirnonton, Dean Keith. 1984. Genius, Creativig, and Leadership. Cambridge, MA.
: Harvard University Press.

Sirnonton, Dean Keith. 1987. Why Presidents Succeed: A mlitical P§ychology of
Leadership. New Haven: Yale University Press.

Sinclair, Barbara. 1993. Studying Presidential Leadership. In Researching the

Presidency: Vital Questions, New Approaches, edited by George C. Edwards 111,
et. al., 203-232. Pittsburgh and London: University of Pittsburgh Press.

Stewart, Louis H. 1977. Birth Order and Political Leadership. In A Psychological
Examination of Political Leaders edited by Margaret G. Hermann, 206-236.
New York: Free Press.

 

Stogdill, Ralph Melvin. 1981. Stogdill’s handbook of Leadership: A Survey of Theory

and research. Revised and expanded edition. New York: The Free Press.

Suedfeld, Peter and Otto Rank. 1976. Revolutionary Leaders: Long-term Success as
a Function of Changes in Conceptual Complexity. Journal of Personaliry and
Social Psychology, 34 (August): 169-178.

Takane, Masa’aki. 1981. The Political Elite in Japan. Berkeley: University of
California.

Taylor, Charles Lewis and David A. Jodice, eds. 1983. World Handbook of Political
and Social Indicators 111: 1948-1982. New Haven, Conn: Yale University.

Taylor, M. and V. M. Herman. 1971. Party Systems and Government Stability.
American Political Science Review 65:28-37.

 

Thompson, Kenneth W. 1984. The Dilemmas Antinomies of Leadership. In

Leadership: Interdisciplimy Reﬂections, edited by R. S. Khare and David Little.
Lanharn, New York, and London: University Press of America Inc.

Thompson, Kenneth W., Hans Morgenthau, and David Hein, eds. 1983. Essays on
Leadership: Compgative Insights. Lanharn: University of America.

205

Topornin, Boris N. 1990. On the classiﬁcation of Socialist political system. In
Contempprg Political Systems: Classiﬁcations and Tymlogies, edited by Anton

Bebler and Jim Seroke. Boulder: L. Rienner.
Trtunan, David B. 1951. The Governmental Process. New York.

Tsuritani, Takesugu and Jack B. Gabbert, eds. 1992. Chief Executives: National

Political Leadership in the United States, Mexico, Great Britain, and Japan.
Pullman, Washington: Washington State University.

Tucker, Robert C. 1977. Personality and Political Leadership. Political Science
Quarterly, Vol.92, No.3 (Fall): 383-393.

Tucker, Robert C. 1981. Politics as Leadership in Political Leadership in Democratic
Societies. Columbia: University of Missouri Press.

Verba, Sidney. 1961. Small Groups and Political Behavior. Princeton University Press.

Vemoff, Edward, and Rirna Shore. 1987. The International Dictiom of 20th Cenrm
Biography. New York : New American Library.

Weisberg, Herbert F. and Jerrold G. Rusk. 1970. Dimensions of Candidate Evaluation.
American Political Science Review Vol.64, No.4 (December): 1167-1185.

 

Weisberg, Herbert F. 1978. Evaluating Theories of Congressional Roll-Call Voting.
American Journal of Political Science 22: 554-577.

 

Welsh, William A. 1970. Methodological Problems in the Study of Political
Leadership in Latin America. Latin American Research Review V (Fall): 3-34.

 

Who’s Who in Australasia and the Far East. 1989. 1st ed. International Biographical
Centre.

Who’s Who in the World. 1971/72-1992. Chicago, Marquis Who’s Who, Inc.

Willner, Ann Ruth. 1968. The Smllbinders: Qharismatic Political Leadership. New
Haven: Yale University Press.

Wiseman, John A. 1991. Political Leﬂers in Black Africa : A Biographical Dictiom

of the Major Politicians Since Indegndence. Edward Elgar Publishing
Company.

Wriggins, W. Howard. 1969. The Ruler’s Immrative: Strategies for Political Survival
in Asia and Africa. Southern Asian Institute Series.

206

Zimmerman, E. 1979. Crisis and Crisis Outcomes: Towards a New Synthetic
Approach. Euroman Journal of Political Research, 7 (March): 67-115.

   

"711 11111111111111111111“