Codebook:
Kapstein and Converse Dataset of Young Democracies, 1960-2004
DEFINITION OF DEMOCRACY ______________________________________________ 2
VARIABLES INCLUDED IN DATASET ________________________________________ 5
Identifier Variables _________________________________________________________ 5
Political Variables __________________________________________________________ 5
Other Variables ____________________________________________________________ 6
NOTES ON COUNTRY DEFINITIONS _________________________________________ 6
REFERENCES ______________________________________________________________ 7
Definition of Democracy
Classifying governments as democratic or non-democratic, as well as identifying episodes of
political change as democratizations is fraught with theoretical and practical pitfalls. We have
attempted to avoid engaging with this particular issue by first identifying democratizations, and
then classifying the regime prior to the democratization as undemocratic and the resulting regime
as a new democracy. Consequently, a more accurate term for the countries in our dataset might
be “newly democratized” countries, rather than “democracies.”
In building our democratization dataset, we have relied on the Polity IV dataset on political
regime characteristics and transitions (Marshall and Jaggers 2005). We define
“democratizations” based on the "Major Democratic Transitions" in the Polity dataset. Such
transitions involve a six-point or greater increase (e.g. from “-3” to “+3”) in the overall Polity
score over a period of three or fewer years. This definition of democratization has been used by
other researchers, for example Rodrik and Wacziarg (2005) and Rigobon and Rodrik (2004).
Whereas the Polity data takes pains to identify and flag transition periods between regimes, our
analysis considers democratizations to have taken place only when the transitions have actually
“finished” and led to a democratic regime. For example, Polity views Mexico as undergoing a
gradual democratic transition from 1994 to 1997, with each of those four years considered to
have contained a “Major Democratic Transition.” Our analysis classifies 1997 as the first year of
the new democracy, since it was in that year the new, more democratic system was, at least
according to Polity, fully in place.
We classify countries as undergoing reversals, i.e. ceasing to be democracies, when they
experience what the Polity dataset terms an “adverse regime transition,” an anti-democratic
“revolutionary transformation in the mode of governance” (Marshall and Jaggers 2005, 35).
We freely confess that, in theory, our methodology is problematic for at least two reasons, which
fortunately do not arise in practice with the data. First, our approach allows for the possibility of
two successive democratizations with no intervening reversal or transition, which is theoretically
problematic, since the academic literature on democratization generally conceptualizes it as an
“event,” with subsequent increases in the quality of democracy characterized as steps toward
consolidation. However, this does not actually occur in the Polity data after 1960. In two
cases—Guatemala in the 1970s and Thailand in 1991—Polity data identifies on or more
"negative regime changes" (as opposed to a more serious "adverse regime transition") between
episodes of democratization. In these situations, the first of these setbacks to democracy was
coded as a democratic reversal in our dataset.
A second and related issue stems from focusing on quantitative changes rather than qualitative
levels to separate democracies from non-democracies. In theory, a six-point “Major Democratic
Transition” could bring a country from a higher to a lower authoritarian score (e.g. “-10” to “4”), resulting in an undemocratic government being mistakenly classified as having become
democratic. Again, this does not in fact occur in the data set, and all regime changes classified as
2
democratizations in this study produced a positive Polity score. Four have a score of one (Sierra
Leone 1968, Ethiopia 1995, Pakistan 1962, Cambodia 1993), eight have a score of two, seven
with a score of three, nine with a score of four, and twelve with a score of five.
A further methodological issue arises regarding the treatment of newly independent states. Some
of the most widely cited empirical studies of the effects of democracy on economic policy and
performance did not incorporate data from the newly independent nations of the former Soviet
bloc, leading scholars to question the robustness of the findings. For those studies produced since
1990, this is partly because no convention exists among scholars regarding the political regime
classification of these states. Presumably, one could investigate the system of government that
prevailed in each newly independent country prior to independence, then judge whether
independence marked a transition to democracy or the consolidation of already extant democratic
institutions. In this case, many of the former Soviet republics would be classified as new
democracies because the USSR was clearly not a democracy. The situation would be less clear
in the case of, for example, British colonies which had some sort of representative assembly,
albeit with limited decision-making power, prior to independence. For the sake of completeness
and consistency, not to mention simplicity, our study classifies all newly independent countries
with positive Polity scores as new democracies. This results in the inclusion of 32 additional
new democracies, including 17 countries that were previously part of the Soviet Union,
Yugoslavia, or Czechoslovakia. Our dataset does not include former Soviet republics
Turkmenistan, Tajikistan, Kyrgyzstan, Uzbekistan, and Kazakhstan, which have never attained a
positive Polity score.
This methodology identifies 123 cases of democratization during the period 1960 to 2004. Table
1 below summarizes the distribution of these democratizations across region and decade, while
table 2 indicates the outcome of these democratizations.
Table 1: Democratizations by Region and Decade
1960s 1970s 1980s 1990s
Total
After
2000
Total
26
20
17
51
9
123
Latin America
6
3
11
5
1
26
Western Europe
Eastern Europe
1
0
3
0
0
0
0
19
0
2
4
21
Sub-Saharan Africa
15
6
2
18
5
46
Middle East-N. Africa
0
1
1
1
0
3
Asia
4
7
3
8
1
23
Source: PolityIV, Author's Calculations
3
Table 2: Democratizations by Region and Outcome
Sustained
Reversed
67
56
Latin America
17
9
Western Europe
Eastern Europe
3
19
1
2
Sub-Saharan Africa
17
29
Middle East-N. Africa
1
2
Asia
10
13
Total
Source: PolityIV, Author's Calculations
4
Variables Included in Dataset
IDENTIFIER VARIABLES
id: Unique identifier for each country-year. First two digits identify the country. Last four digits
identify the year.
caseid: identifier code for each episode of democracy. First two digits identify the country. Last
four digits indicate the year of democratization.
ccode: Two digit country code.
country: Country name.
year: year of the observation.
POLITICAL VARIABLES
demage: Years since democratization. Year of democratization is one. Blank in non-democratic
years
demduration: Final length of that particular episode of democracy, as of 2004. Blank in nondemocratic years.
lastyear: 1 in the last year of an episode of democracy that was reversed, 0 otherwise during
democratic periods. Blank in non-democratic years.
pres: 1 if regime is presidential, 0 if parliamentary. Blank when no data available and in years of
non-democracy. Combines PresWB and PresP to get coverage from 1960 to 2004. Where PresP
= 2, PresWB classification used. Where the two sources conflicted, continuity was used to
determine which was appropriat.
priordem: Indicates the number of democratic episodes the country has experienced since 1800,
including the current episode. Blank in non-democratic years.
reversed: 1 in all years of democratic episode that ended in reversal. 0 in all years of a
democratic episode that, as of 2004, had not been reversed. Blank in non-democratic years.
xconst: PolityIV variable EXCONST.
xconstfall: 1 in all years of an episode of democracy following a loosening of constraints on the
executive. 0 if no fall in the executive constraints score has occurred previously during that
particular episode of democracy.
xconsthi: XConst is 5 or greater.
5
xconsthi2: XConst is 7 or greater.
xconstfall: 1 in all years of an episode of democracy following a tightening of constraints on the
executive. 0 if no fall in the executive constraints score has occurred previously during that
particular episode of democracy.
OTHER VARIABLES
liberal: 1 if classified as liberalized economy by Horn Welch and Wacziarg (2003) in their
updating of the Sachs-Warner index. 0 if classified as not liberalized
liberalize: 1 in year when liberal goes from 0 to 1
libdem: 1 if economic liberalization, as indicated by a change in Horn Welch and Wacziarg’s
(2003) update of the Sachs-Warner index, has previously occurred during the particular episode
of democracy.
postcol: 1 in all years of democratic episodes which were immediately preceded by colonial rule.
0 if democratic episode was not postcolonial, and blank in years of no democracy.
postcom: 1 in all years of democratic episodes which were immediately preceded by communist
rule. 0 if democratic episode was not postcolonial, and blank in years of no democracy.
region: Indicates region:
0 North America
1 Latin America and Caribbean
2 Western Europe
3 Eastern Europe
4 Sub-Saharan Africa
5 North Africa-Middle East
6 Asia and Oceania
Notes on Country Definitions
Czech Republic/Czechoslovakia: Single code (315) for Czechoslovakia and Czech Republic,
with single democratization.
Yugoslavia: Yugoslavia classified as same country before and after secession of member
republics. Single code 345.
Russia: Single code (365) for USSR and Russia.
Ethiopia: Coded as the same country before and after secession of Eritrea. Single code 530.
Pakistan: Codes as the same country both before and after secession of Bangladesh. Single code
770
6
References
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain
Wacziarg. 2003. Fractionalization. Journal of Economic Growth 8: 155-194.
Marshall, Monty G. and Keith Jaggers. 2002. Polity IV Project: Political Regime
Characteristics and Transactions, 1800-2002. University of Maryland.
Przeworski, Adam, Michael E. Alvarez, Jose Antonio Cheibub and Fernando Limongi. 2000.
Democracy and Development: Political Institutions and Well-being in the World, 1950–
1990. (New York: Cambridge University Press).
Rigobon, Roberto and Dani Rodrik. 2004. Rule of Law, Democracy, Openness, and Income:
Estimating the Interrelationships. NBER Working Paper 10750. Cambridge, MA:
National Bureau of Economic Research.
Rodrik, Dani and Romain Wacziarg. 2005. Do Democratic Transitions Produce Bad Economic
Outcomes?” American Economic Review 95( 2): 50-55.
Wacziarg, Romain T. and Horn Welch, Karen. 2003. Trade Liberalization and Growth: New
Evidence. NBER Working Paper No. W10152. .
7