1 The reader’s guide to the dataset (and codebook)

1.1 Structure of the codebook

This codebook is organized as follows:

PART I: outlines case identifiers.

PART II: includes contested states’ lifecycle indicators, such as those capturing their emergence, livelihood, and death. This part also includes similar indicators on the respected parent states.

PART III: includes indicators that capture various aspects of the stateness of contested states, such as their domestic institutional and bureaucratic set-up, among others. This part also includes similar indicators on the respected parent states.

PART IV: includes indicators on contested states’ territorial characteristics, such as their size, actual area under control, border length, territorial contiguity, and others. This part also includes similar indicators on the respected parent states.

PART V: includes demographic indicators of contested states and their respected parent states.

PART VI: includes economic indicators of the respected parent states.

PART VII: includes indicators on military capacities of both contested states and their respected parent states.

PART VIII: includes several conflict-related indicators where either contested states or their respected parent states have been engaged in.

PART IX: includes indicators on various forms and types of external support that contested states have received

PART X: includes indicators on various international missions or other forms of foreign presence in contested states.

PART XI: includes indicators on contested states’ international legitimacy.

PART XII: includes indicators on the attention paid to contested states by the international community, namely the ways these entities have featured in different international fora, such as the UN General Assembly, UN Security Council, and several other UN-related agencies.

PART XIII: includes some other indicators.

PART XIV: outlines the transformed versions of some of the existing indicators mentioned in the previous parts. Transformed data, such as categorized, normalized, and/or rescaled versions of the continuous indicators, are provided to support some research contexts where such transformed indicators can be more useful. The transformed indicators can also be useful when building indicators of a higher order of abstraction (higher level of aggregation) by both arithmetic addition, subtraction, division, or multiplication, and set addition and multiplication.

PART XV: outlines several indicators of a higher order of abstraction that are aggregated by combining some of the existing indicators mentioned in the previous part to capture certain conceptual phenomena.

1.2 Indicator lots

All indicators belong to different lots, which help the reader understand the nature of each indicator. Each indicator of the dataset belongs to only one of the following lots:

LOT (A): Indicators allocated to this lot refer to all authentic (original) indicators on contested states that have been gathered from primary and secondary sources, which, to the knowledge of the author, cannot be found in any of the existing datasets.

LOT (B): Indicators allocated to this lot refer to all indicators which, as such, do not exist in any of the existing datasets, but are build through different recoding procedures of the empirical evidence available in existing datasets. To illustrate with an example, the UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) provides vast empirical evidence on external support to, among others, contested states. Moreover, the UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) provides such dichotomous indicators as (external_type_X) - whether an external supporter provided troops to an entity, or (external_type_W) - whether an external supporter provided weapons to an entity, and a score of other indicators marking other forms of external support. These UCDP indicators, however, do not tell us much about the type of actors providing these forms of support, for example, if troops were provided by a state or non-state actor, or from a neighbor or a Great Power. Yet, the UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) provides another indicator called (external_name), which provides the names of external supporters providing different forms of external support. I use such indicators to recode the existing empirical evidence provided by the UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011), whereby I use (external_name) to extract the type of actor (i) state; (ii) non-state; (iii) neighbor; (iv) Great Power, and so on, as well as all the indicators such as (external_type_X), (external_type_W), and others, to recode them into my indicators such as (troopsN) - whether troops have been provided by a neighboring state or not, or (troopsGP) - whether troops have been provided by a Great Power, and so on. In sum, indicators that belong to Lot (B) are the ones that do not exist as such in any of the existing datasets but are build by recoding the empirical evidence provided in the existing datasets to make (better) use of such available empirical evidence from the existing datasets.

Lot (C): Indicators allocated to this lot are the ones that are taken from existing datasets with no substantial changes made. No substantial changes here refers to such changes as adding and/or imputing some missing values or adding a few cases for which the empirical evidence has been registered by following all the data coding procedures provided in the codebooks of these existing datasets. If the data format, such as date format, and other similar parameters are changed, that the indicator will be marked under Lot (B).

Lot (D) Indicators allocated to this lot include those that are aggregated from indicators belonging to lots (A), (B), and/or (C). For example, Lot (D) includes indicators derived by arithmetic division, such as population per capita, GDP per capita, military expenditure per capita, and so on. This lot also includes indicators which are created by cumulative scale, such as cumulative days at war, or cumulative years of existence, and alike. Furthermore, Lot (D) includes those that are arrived at by arithmetic addition, such as one of the alternatives of the measurement model of external support whereby each form of external support is added to capture the intensity of external support. Arithmetic multiplication is also used with some aggregated indicators. All aggregation procedures are described in detail as part of indicator descriptions whenever such indicators are generated through various aggregation models.

1.3 Rules for naming indicators

Indicators are named according to the following rules:

Rule 1: All indicators are presented in parenthesis in italicized form, such as (namecs) for the name of contested states, (namep) for the name of parent states, or (controlcs) for the approximate territorial area controlled by the contested state, and so on.

Rule 2: All indicators belonging to lots (A), (B), and (D) are named by the author. The indicators belonging to Lot (C), whenever possible, preserve their original names, which they hold in the datasets from which they are borrowed. This is done to preserve consistency across datasets for validity and comparability purposes.

Rule 3: For all indicators that are transformed, suffixes are added to their original names to indicate some information with regards to their transformation. Continuous indicators that are transformed into a categorical (usually ordinal) scale, the following suffix is added to the original indicator name: "_x_cat“, where”x" indicates the number of categories in which the indicator is divided, while “cat” indicates that the indicator has been transformed into a categorical scale. For example, (factions_4_cat) represents an ordinal scale of 4 categories (1 to 4) of the original indicator (factions) with values from 1 to 18. Continuous indicators that are normalized and/or rescaled into another continuous indicator, the following suffix is added to the original indicator name: "_x_y“, where”x" indicates whether the indicator has been rescaled, while, while “y” indicates whether and the form through which the indicator has been normalized. For example, (milexcap_r_croot) represents a rescaled and cube root normalized continuous version (with transformed values between 0 and 5) of the original continuous indicator (milexcap) with original values between 0 and 2331.

Rule 4: For some indicators, usually those of higher order of abstraction, that are aggregated by combining some of the existing indicators, alternative measurement models are presented. For every alternative version of the same indicator, a suffix of the form "_alt_x" is added to the original name of the indicator. For example, the original indicator “extsupport”, representing an aggregated indicator of external support to contested states has three other alternative measurements, which subsequently are named “extsupport_alt_1”, “extsupport_alt_2”, “extsupport_alt_3”, and so on

1.4 Data sources and formats

Format: Textual.

Type: Categorical.

Scale: Nominal.

Source: Gleditsch and Ward (1999).

This indicator registers the yearly age difference between the parent state and the contested state. The (diffy) is extracted from (diffd) by the following calculation:

\[diffy = \frac {diffd}{365}\]

Note: For each observed year of each contested state, there is a fraction of year added to the age difference between the contested state and the parent state. This is because of the software’s (R) differences in date calculations, which, because of the leap years, sometimes it takes the year to be 364 days long and at other times 365 days.

Format: Numerical.

Type: Continuous.

Scale: Interval.

Source: Calculation from (diffd).

3.18 Contested state outcome (outcome) | (A)

This indicator registers the two main (life) outcomes of contested states for each observed year, and it takes one of the following two values:

1 – If the contested state falls under the category of survival. These are cases that never went back under their parent states’ control; they either continue surviving without losing any of the necessary dimensions of the concept of contested state, or they survived by already managing to upgrade their status to full members of the United Nations.

0 – If the contested state falls under the category of death (not survival). These are cases that went back under their parent state’s control, either peacefully or violently.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: See “is it a contested state?” part of the discussion in the “Contested State Empirical-Procedural Applicability” appendix.

3.19 Contested state sub-outcome (suboutcome) | (A)

This indicator registers the sub-outcomes of contested states within the main two outcomes in (outcome) for each observed year. The (suboutcome) captures the ways the contested states survived or died.

1 – If the contested state continues surviving. These are the cases that continue to survive as contested states; meaning, they did not fall under any of the other sub-outcomes. This is a sub-outcome of the (outcome [1]).

2 – If the contested state survived. These are the cases that eventually “graduated” to become full members of the United Nations, and did not fall under any of the other sub-outcomes. This is a sub-outcome of the (outcome [1]).

3 – If the contested state died through a peaceful reintegration into the parent state, usually through some negotiations process. This is a sub-outcome of the (outcome [0]).

4 – If the contested state died through a violent reintegration back into the parent state, usually through an invasion by the parent state. This is a sub-outcome of the (outcome [0]).

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: See in the discussion in the “Contested State Empirical-Procedural Applicability” appendix.

3.20 Maturity of the parent state (newstate)| (A)

This indicator registers the age of the parent state for every case-year observation of contested states, and it takes the following four values:

3 – If the parent state is 1 year old or younger.

2 – If the parent state is between 1 to 2 (inclusive) years old.

1 – If the parent state is between 2 to 3 (inclusive) years old.

0 – If the parent state is older than 3 years.

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Considerations from (unp).

4 PART III: Stateness indicators

4.1 Contested state passports (passportcs) | (A)

This indicator registers whether contested states have issued their own passports during each observed year, and can take one of the following two values:

1 – If the contested state has issued its own passports to its citizens. To take this value, a contested state must cover a substantial number of its population with passports. To illustrate, at some point, Chechnya had issued self-styled passports to a few hundred of its high-ranking officials. Such cases are excluded as having issued their own passports, as they do not cover a substantial number of their populations.

0 – If the contested state has not issued its own passports to its citizens.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.2 Parent state passports (passportp) | (A)

This indicator registers whether contested states have used the parent state’s passports during each observed year, and it can take one of the following two values:

1 – If the contested state has used its parent state’s passports. For the contested state to take this value, a parent state must cover a substantial amount of the contested state’s population with passports. For example, if only some individuals from the contested state are reported to have in their possession the parent state’s passports, the case does not count for this category.

0 – If the contested state has not used the parent state’s passports.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.3 Third-party passports (passportes) | (A)

This indicator registers whether contested states have used any third-party state’s (usually, but not always, external supporter’s) passports for each observed year. The (passportes) can take one of the following two values:

1 – If the contested state has useda third-party state’s passports. For this category, a third-party state needs to cover a substantial amount of the contested state’s population with its passports. For instance, the issuance of Russian passports to the populations of Lugansk and Donetsk Republics has not been significant enough (at least not until 2017) to put these cases in this category of (passportes). The populations of Donetsk and Lugansk Republics continued to use Ukrainian (parent state’s) passports in substantial amounts.

0 – If the contested state has not used a third-party state’s passports.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.4 Contested state central bank (cbank) | (A)

This indicator registers whether contested states have had an established central bank of their own for each observed year. The payment authority as an institution, such as Kosovo’s case before 2010, or Palestine’s case throughout its existence, is not counted as a central bank. This indicator can take one of the following two values:

1 – If the contested state had a central bank.

0 – If the contested state had no central bank.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.5 Contested state currency (currencycs) | (A)

This indicator registers whether contested states have issued their own currency for each observed year. The (currencycs) does not include cases which have issued commemorative notes or coins, such as the cases of Abkhazia, South Ossetia, Western Sahara, etc., and can take one of the following two values:

1 – If the contested state has issued its own currency. This category also includes cases that have pegged their currency to a third-party state (usually an external supporter), so long as they had their own issued currency in circulation.

0 – If the contested state has not issued its own currency.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.6 Parent state currency (currencyp) | (A)

This indicator registers whether contested states have used their parent state’s currency in each observed year, and can take one of the following two values:

1 – If the contested state has used the parent state’s currency.

0 – If the contested state has not the used parent state’s currency.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.7 Third-party currency (currencyes) | (A)

This indicator registers whether the contested state has used any third party state’s (usually external supporter’s but not always necessarily) currency in each observed year, and can take one of the following two values:

1 – If the contested state has used a third-party state’s currency

0 – If the contested state has not used a third-party state’s currency

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.8 Control over the capital city (capitalcontrol) | (A)

This indicator registers whether contested states maintained control over their claimed capital city during each observed year, and can take one of the following two values:

1 – If the contested maintained control over its capital city.

0 – If the contested maintained had control over its capital city.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.9 Contested state ministry of foreign affairs (mfa) | (A)

This Indicator registers whether contested states had an established Ministry of Foreign Affairs (MFA) in each observed year, and can take one of the following two values:

1 – If the contested state had an MFA.

0 – If the contested state did not have an MFA.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

4.10 State-building efforts by the contested state (dfsbuild) | (C)

This indicator represents Florea’s (2014) measurements of the degree of state-building in contested states. For cases and/or years that do not feature in Florea (2014), but feature in this dataset, I have used additional sources to capture the (dfsbuild). When doing so, I used Florea’s (2014) description of the indicator in his supplementary material as a guide for the operationalization and measurement of the collected data. The (dfsbuild) can take one of the following four values:

1 – Low degree of state-building efforts.

2 – Moderate degree of state-building efforts.

3 – High degree of state-building efforts.

4 – Very high degree of state-building efforts.

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Florea (2014), Lexis Nexis, and Factiva.

4.11 Number of factions in contested states (factions) | (C)

The number of factions in each contested state is captured from Cunningham’s (2013) data on factions in self-determination (SD) movements. According to Cunningham’s (2013) conceptualization, to be identified as a faction in SD, two criteria need to be met. First, the faction must represent the SD (in my case, the contested state); and second, a faction must make demands on behalf of the SD (in my case, the contested state). For more detailed information on the procedures of how the number of factions is captured, see Cunningham’s (2013) supplementary material.

Cunningham’s (2013) dataset on factions covers the period between 1960 - 2005. I use additional sources to capture data on contested states’ factions before 1960 and after 2005. While collecting data from additional sources, I attempt, for consistency purposes, to maintain the same concept and procedure provided in Cunningham (2013) to identify factions for each case-year of contested states that Cunningham (2013) does not cover. The cases which are not included in Cunningham (2013) at all, but are included in my dataset are the following: Donetsk Republic (2014-2017); Hyderabad (1947-1948); Islamic State (2014-2017); Kosovo (2018-2017); Republic of Mahabad (1945-1946); Lugansk Republic (2014-2017); Rwenzururu Kingdom (1963-1982); South Moluccas (1950); and Taiwan (1971-2017).

The (factions) indicator takes any full number between 1 (least factionalized contested state) to 18 (most factionalized contested state) in any given year.

This indicator captures the ratio of the actual area controlled by the contested state against the actual area claimed in each observed year. The (areacontrolcs) is calculated from (controlcs) and (areacs) as follows:

\[areacontrolcs = \frac {controlcs}{areacs}\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (controlcs) and (areacs).

5.6 Ratio of area under the control of parent states (areacontrolp) | (D)

This indicator captures the ratio of the actual area controlled by the parent states against its official area in each observed year. The (areacontrolp) is calculated as follows:

\[areacontrolp = \frac {controlp}{areap}\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (controlp) and (areap).

5.7 Ratio of the official size of the area of contested states against their parent states (arearatio) | (D)

This indicator captures the ratio between the contested state and parent state official areas in each observed year. The (arearatio) is derived as follows:

\[arearatio = \frac {areacs}{areap}\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculations from (areacs) and (areap).

5.8 Ratio of the controlled size of the area by contested states against their parent states (controlratio) | (D)

This indicator captures the ratio between the contested state’s and the parent state’s actual are under their control in each observed year. The (controlratio) is derived as follows:

\[controlratio = \frac {controlcs}{controlp}\]

Format: Numerical.

Type: Continuous

Scale: Ratio

Source: Calculation from (controlcs) and (controlp)

5.9 Contested states’ access to sea (seacs) | (A)

This indicator registers whether contested states have had access to sea during each observed year. The (seacs) captures the fact of having access to the sea and not whether the officially claimed territory of the contested state has had access to the sea. For example, even though the map of Biafra shows that the entity has had access to the sea, Biafra is registered as having had access to the sea only when it actually did so, namely in 1967 when it emerged. However, between 1968 to 1970 (until its death as a contested state), Biafra is registered as not having had access to the sea given the Nigerian government control over the coastal areas of Biafra during this period. The (seacs) takes one of the following two values:

1 – If the contested state has had access to the sea.

0 – If the contested state has not had access to the sea.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous

Source: To be provided upon request

5.10 Parent states’ access to sea (seap) | (A)

This indicator registers whether parent states have had access to the sea for each observed year. Just like with the same indicator for contested states, the (seap) registers the parent states’ actual access to the sea, and can take one of the following two values:

1 – If the parent state has had access to the sea.

0 – If the parent state does has not had access to the sea.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.11 Contested states as islands (islandcs) | (A)

This indicator registers whether contested states have been islands during each observed year. The (islandcs) captures the fact of being an island, and not whether the officially claimed territory is an island. For example, despite that the map of Bougainville shows that the entity is an island, it is registered as being an island only in 1990 when the Papua New Guinean (PNG) Government forces were expelled from the island completely, and the island fell under the control of Bougainville Revolutionary Army (BRA). However, between 1991 to 2001 (until its death as a contested state), Bougainville is registered as a non-island, given that the PNG Government troops entered and controlled between 10 to 70 percent of the island depending on the period. This means that the contested state of Bougainville bordered its parent state on the island. The (islandcs) takes one of the following two values:

1 – If the contested state has been an island.

0 – If the contested state has not been an island.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.12 Parent states as islands (islandp) | (A)

This indicator registers whether parent states have been islands for each observed year, and it takes one of the following two values:

1 – If the parent state has been an island.

0 – If the parent state has not been an island.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.13 Contested states’ territorial contiguity (contigcs) | (A)

This indicator registers whether contested states have had contiguous territories during each observed year. Just like with other indicators of territorial characteristics, the (contigcs) registeres the fact of contiguity, and not the official territorial contiguity, and takes one of the following two values:

1 – If the contested state has had a contiguous territory.

0 – If the contested state has not had a contiguous territory.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.14 Parent states’ territorial contiguity (contigp) | (A)

This indicator registers whether parent states have had contiguous territories during each observed year, and it takes one of the following two values:

1 – If the parent state has had a contiguous territory.

0 – If the parent state has not had a contiguous territory.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.15 Contested states’ land border with their parent states (borderwp) | (A)

This indicator captures whether contested states have had shared land borders with their respective parent states for each observed year. The (borderwp) does not include maritime borders, i.e., cases in which contested states and parent states share a certain distance of sea/ocean water among them. Some of these cases include Taiwan v.s. China, Anjouan v.s. Comoros, among others. However, if contested states and parent states have rivers that divide them (such as Transnistria v.s. Moldova), they will be considered as sharing a land border. The (borderwp) takes one of the following two values:

1 – If the contested state shared a land border with the parent state

0 – If the contested state has not shared a land border with the parent state

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.16 Contested states’ approximate land border length with their parent states (blengthwp) | (A)

This indicator captures the approximate length in km of contested states’ land borders with their respective parent states for each observed year.

Format: Numerical (km).

Type: Continuous.

Scale: Ratio.

Source: To be provided upon request

5.17 Contested states’ land borders with entities other than their parent states (bordero) | (A)

This indicator registers whether contested states have had land borders with entities other than their respective parent states in each observed year. Even if contested states do not border another UN member state, but border another contested state in the territory of their respective parent states, they are registered as having had a shared land border with an entity other than their respective parent states. Some examples include the Donetsk Republic, which borders the Lugansk Republic, but no other state or entity except for its parent state (Ukraine), and South Kasai, which bordered Katanga – another contested state in D.R Congo - but no other entity except for its parent (D.R. Congo). Hyderabad, on the other hand, is an example of sharing no border with entities or states other than its parent state – India. Moreover, just like with the other indicators, it is worth noting that (bordero) registers the fact of sharing a border with another state or entity, and not the officially claimed version. For example, although Palestine’s map shows that it shares a border with entities other than the parent state, in fact, it does not, because of Israel’s control of Palestine’s borders in the West Bank, and the blockade of Gaza. This indicator takes one of the following two values:

1 – If the contested state had a shared land border with an entity other than the parent state.

0 – If the contested state has not had a shared land border with an entity other than the parent state.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

5.18 Contested states’ land border with external supporters (borderwes) | (A)

This indicator captures whether contested states have had shared land borders with their respective external supporters (including non-state actor supporters) for each observed year. This does not include maritime borders – i.e., in case the contested state or the external supporter is an island. The (borderwes) takes one of the following two values:

1 – If the contested state shared a land border with the external supporter.

0 – If the contested state has not shared a land border with the external supporter.

This indicator registers the ratio between a contested state’s and its respective parent state’s real populations during each observed year. The (popratio) is calculated from (popcs) and (poppwocs) as follows

\[popratio = \frac {popcs}{poppwocs}\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (popcs) and (poppwocs)

7 PART VI Economic indicators

Economic indicators are registered only for the parent states. Economic data on contested states are difficult to secure for the overwhelming majority of the cases. When such data exist, they do so only for a fraction of the time of contested states’ existence - this is especially the case for historical cases of contested states. Also, often, economic data on contested states tend to be highly unreliable as usually little information is provided on such aspects as the currency which the numbers speak off, whether they are based on current or constant rates, which makes any comparative analysis impossible, if not fruitless. The economic indicators that are recorded here for the parent states can be useful when researching contested states, as they can be used to measure parent state’s strength and weakness, growth, and decline, and other related aspects.

7.1 Parent state’s GDP per capita (cgdppc) | (C)

This indicator captures parent states’ real GDP per capita over the observed years in the dataset. The bulk of the data on real GDP per capita is primarily secured from the Maddison Project Database (MPD) version 2018 (Bolt et al. 2018). As per MPD’s description, the (cgdppc) represents the real GDP per capita in 2011 USD, which is suitable for cross-country income comparisons (Bolt et al. 2018). This is, therefore, a measure that corrects for inflation over time and makes GDP data also comparable over time (Bolt et al. 2018). The MPD dataset has an advantage as it is a collaborative economic data collection project and, most importantly, includes historical data. Especially data after 1945 are abundant for all the countries. The MPD dataset, nevertheless, has some missing data. I follow different strategies to find the rough estimates of the missing data in MPD dataset, depending on the nature of the missing data, as follows:

There are several parent states which the MPD does not make data available for, or for the required years in my dataset. These are Papua New Guinea (Bougainville’s parent state), Somalia (Somaliland’s parent state), and Iran (Kurdish Republic of Mahabad’s parent state). For many of these cases, the MPD does provide data, but they do not exist for the specific years that my dataset covers. I use the best available alternative sources and attempt to save for the consistency with the existing MPD data.

For Papua New Guinea, I use the World Bank’s (2019a, see Country: Papua New Guinea) data on GDP per capita (constant 2011 international USD) for the required years to cover the entire case (1990-2001).

For Somalia, the World Bank does not possess GDP constant measures; thus, I use the World Bank’s (2019b, see Country: Somalia) data on current GDP per capita (USD), which only covers the period (2013-2017). I include the CPI data in the calculation to correct for inflation, so to get close rough estimates of GDP per capita on constant Data on Somalia’s GDP generally are absent for the period before 2013, among many other sources. I extrapolate Somalia’s data for the period 1991-2012 by taking three steps that could lead to the best rough estimates for the missing data. First, given that 2013 is the earliest year for which reliable GDP data for Somalia exist, I extract the average real GDP per capita for seven countries of the extended Horn of Africa region (Djibouti, Eritrea, Ethiopia, Kenya, Sudan, South Sudan, and Uganda) for the period 1991-2013 for which data is available from World Bank (2019a). These seven countries are the ones that constitute an economically integrated region in Africa (together with Somalia) and are all members of the Intergovernmental Authority on Development (IGAD). It is often the case that the economies of members of IGAD are analyzed together as they share common resources and industrial and labor base, and enjoy other similarities (for more see: IGAD 2019). Second, I extract the rate of average real GDP per capita of the countries of this region for each year between 1991 and 2012 against the 2013 data. For example, the average real GDP per capita of these countries in 2013 was 2347.46 USD; in 2012, it was 2217.77 USD; in 2011, it was 2363.38 USD, and so on (World Bank 2019a). This means that the rate of the average real GDP per capita of the countries of the region against 2013 data for 2012 was 0.94; for 2011 it was 1.01, and so on. Finally, to extract Somalia’s real GDP per capita for the period 1991-2012, I use the average rates of the region between 1991 and 2012 as they apply to the region’s average 2013 data, and apply them against Somalia’s 2013 real GDP per capita.

For Iran, I use real GDP per capita data for the period 1945-1946 from Esfahani and Pesaran (2009, see: Figure 1, p.181).

Finally, for the 2017 data which are not available in the latest MPD dataset, I take two steps. First, I obtain the annual GDP per capita growth rate for 2017 from the World Bank’s (2019c) GDP per capita annual growth rate for each of the cases of parent states that feature in 2017 in my dataset. Second, I add the 2017 growth rate to the MPD’s 2016 data to get the real GDP per capita for 2017. I do so to keep the consistency of the MPD data that I use and, thus, make it suitable for cross-case comparison, but also for comparison over time. The World Bank (2019c) does not possess the GDP per capita annual growth rate for Syria, but I apply a 10% decline which is a general trend that Syria’s read GDP followed since 2013 up to 2015 (CIA 2019, see the country: Syria, section: Economy). For Islamic State’s parent states (Syria and Iraq), I use the average of the two parent states’ real GDP per capita for each of their observed years.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Maddison Project Database (MPD), version 2018 (Bolt et al. 2018), World Bank (2019b), World Bank (2019c), CIA (2019), Esfahani and Pesaran (2009).

7.2 World average GDP per capita (wgdpcapita) | (D)

This indicator captures the entire world’s average real GDP per capita during each observed year in the dataset. Data for (wgdpcapita) is secured from the Maddison Project Database (MPD) version 2018 (Bolt et al. 2018). It is calculated by adding the (cgdppc) for each country in each year. This can be useful to examine how parent states stand in each year against the world’s average economy per capita over the years of the existence of contested states.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (cgdppc).

7.3 Parent state’s GDP per capita quantiles (gdpQuantiles) | (D)

This indicator captures the parent states’ GDP per capita in relation to the world’s GDP per capita in 10 quantiles. For every case-year observation, each of the 10 quantiles is different, with the change of the world’s average GDP per capita over all the years.

This indicator captures the percent of the total population mobilized as active military personnel in each contested state during each observed year. The (troopscscap) is extracted from (troopscs) and (popcs) by the following calculation:

\[troopscscap = \frac {troopscs}{popcs} * 100\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (troopscs) and (popcs).

8.4 Percent of population mobilized in parent states (troopspcap) | (D)

This indicator captures the percent of the total population mobilized as active military personnel in each parent state during each observed year. The (troopspcap) is extracted from (troopsp) and (poppwocs) by the following calculation:

\[troopspcap = \frac {troopsp}{poppwocs} * 100\]

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (troopsp) and (poppwocs).

8.5 Military expenditure of parent states (milex) | (A)

This indicator captures the military expenditure of each parent state in USD to the nearest thousand (’000). Data for (milex) are primarily taken from SIPRI (2018), which makes data on military expenditure available for the 1949-2017 period. The SIPRI (2018) data are available in various expressions, including military expenditures in constant USD, in current USD, and in local currencies. I use constant USD (2016) data, as they are adjusted for the effects of inflation, thus enabling cross-country and cross-temporal comparisons to be more reliable based on real military expenditures.

Given that there are some missing values in SIPRI (2018) for some cases and some case-years of parent states, I use the Correlates of War (CoW), National Material Capabilities (NMC), Version 5.0 dataset (Singer, Bremer, and Stuckey 1972; Singer 1988), which, among others, contains data on countries’ military expenditures over 1816-2012 period. The CoW NMC (Singer, Bremer, and Stuckey 1972; Singer 1988) are expressed in current USD, which means that military expenditures are not adjusted for the effects of inflation. To secure the consistency with the SIPRI (2018) data, I convert the CoW NMC (Singer, Bremer, and Stuckey 1972; Singer 1988) military expenditure to constant USD (2016). This means that the CoW NMC (Singer, Bremer, and Stuckey 1972; Singer 1988) military expenditure data are adjusted for the effects of inflation with the 2016 USD value of 1.00. To adjust for inflation in USD terms, I use the Consumer Price Index (CPI) data from US Bureau Labor of Statistics (Bureau of Labor Statistics 2019) for the 1945-2017 period, to calculate the real military expenditures from CoW NMC (Singer, Bremer, and Stuckey 1972; Singer 1988).`

I additionally use military expenditure data provided by World Bank (2019d), which, given that they are provided in current USD terms, I convert them to constant USD (2016), using the CPI data from US Bureau Labor of Statistics (Bureau of Labor Statistics 2019) for the 1945-2017 period. Finally, where there were still missing data for some parent state case-years, I imputed them by taking the average of the earliest and latest case-years, for which the data exist, to get the rough estimates. These namely include data on some case-years for Somalia, Comoros, Syria, and a single case-year for Abkhazia and South Ossetia. Just like with data on real GDP per capita, for Islamic State’s parent states (Syria and Iraq), I use the average of the two parent states’ real military expenditure for each of their observed years.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: SIPRI (2018), CoW NMC (Singer, Bremer, and Stuckey 1972; Singer 1988), World Bank (2019d).

8.6 Per capita military expenditure of parent states (milexcap) | (D)

This indicator captures the per capita military expenditure of each parent state in USD. The (milexcap) is extracted from (milex) and (poppwocs) by the following calculation:

\[milexcap = \frac {milex * 10000}{poppwocs}\] Note: (milex) is multiplied by 1000 to extract the actual full rate of USD per capita expenditures, and not to the nearest thousand (’000)

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Calculation from (milex) and (poppwocs).

9 PART VIII: Conflict indicators

Data on conflicts in which contested states were involved in with their parent states are identified from the Uppsala Conflict Data Program (UCDP) and Centre for the Study of Civil Wars, International Peace Research Institute, Oslo (PRIO) conflict datasets. According to UCDP/PRIO, conflict is defined as a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths in a calendar year (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019). UCDP/PRIO has several different conflict thematic datasets. The ones that I use to identify various indicators for conflicts in which contested states, and their parent states are involved in, I use the following: (i) the UCDP/PRIO Armed Conflict Dataset, Version 19.1 (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019), which covers the period between 1946 and 2017; (ii) the UCDP/PRIO Dyadic Dataset version, Version 19.1 (Harbom, Melander, and Wallensteen 2008; Pettersson, Högbladh, and Öberg 2019), which covers the period between between 1946 and 2017; and (iii) the UCDP/PRIO Conflict Termination Dataset, Version 2-2015 (Kreutz 2010), which covers the period between 1946 and 2017.

While these datasets are organized differently for each conflicting party, I transform UCDP/PRIO’s dataset format to suit the type of time-series dataset I have built for contested states. Below are the detailed descriptions of how the UCDP/PRIO data and indicators are registered in my dataset. Various indicators on conflict are registered for both, contested states and their respective parent states.

9.1 Contested states’ conflict with their parent state (confcs) | (B)

This indicator registers whether contested states have been engaged in any conflict with their respective parent states during each observed year. The (confcs) does not register contested states’ conflicts other than those with their parent state, and takes one of the following two values:

1 – If the contested state has been in conflict with the parent states. This is marked even if the contested state has been engaged in conflict for a fraction of the specified year.

0 – If the contested state has not been engaged in conflict with the parent state.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: UCDP/PRIO Armed Conflict Dataset, Version 19.1 (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019).

9.2 Contested state conflict start (confcst1) | (B)

This indicator captures the date of the calendar year from when a contested state has been in conflict. The (confcst1) registers the actual start date of the conflict during a given calendar year when the contested state existed, even if a contested sate emerged after the conflict start date in that given calendar year. For example, despite that Abkhazia emerged as a contested state in mid-September 1993, Abkhazia’s conflict with Georgia had started much earlier - before 1993. Therefore, the (confcst1) for Abkhazia for 1993 is 1993-01-01, and not 1993-09-15, when it emerged. For contested state case-years for which there was no registered conflict, the fields are marked with NA.

Format: Numerical (date format: yyyy-mm-dd).

Type: Continuous.

Scale: Interval.

Source: UCDP/PRIO Armed Conflict Dataset, Version 19.1 (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019).

9.3 Contested state conflict end (confcst2) | (B)

This indicator captures the date of the calendar year until when a contested state has been in conflict with its parent state. The (confcst2) registers the actual end date of the conflict during a given calendar year when the contested state existed, even if a contested sate ceased to exist before the end of the conflict in that given calendar year. For example, despite that Aceh ceased to exist as a contested state at the end of August 2005, Aceh’s conflict with Indonesia ended after Aceh ceased to exist in that calendar year. Therefore the (confcst2) for Aceh for 2005 is 2005-10-12, when the actual conflict ended, and not 1993-08-31 when the contested state ceased to exist. For contested state case-years for which there was no registered conflict, the fields are marked with NA.

Format: Numerical (date format: yyyy-mm-dd).

Type: Continuous.

Scale: Interval.

Source: UCDP/PRIO Armed Conflict Dataset, Version 19.1 (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019).

9.4 Contested state conflict intensity (csconfint) | (B)

This indicator captures the intensity of the contested state’s conflict with the parent state for each observed year, and it takes one of the following values:

0 – If there was no conflict.

1 – If the conflict was a low-intensity conflict (between 25 and 999 battle deaths).

2 – If the conflict was a high-intensity conflict (1,000 or more battle deaths).

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP/PRIO Armed Conflict Dataset, Version 19.1 (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019).

9.5 Contested state conflict intensity (csconfincompat) | (B)

This indicator captures the contested state’s conflict incompatibility with its parent state during each observed year, and it can take one of the following values:

0 – If there was no conflict.

1 – If the incompatibility was over territory.

2 – If the incompatibility was over government.

3 – If the incompatibility was over both territory and government.

0 – If there was no ceasefire agreement between the contested state and the parent state.

1 – If there was a ceasefire agreement between the contested state and the parent state.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: UCDP/PRIO Conflict Termination Dataset, Version 2-2015 (Kreutz 2010).

9.28 Date of ceasefire agreement (confcfiredate) | (B)

This indicator captures the date when the ceasefire agreement was reached. If there was no ceasefire agreement, the field for the date is marked with NA. For contested state case-years for which there were no registered ceasefire agreements, the fields are marked with NA.

Format: Numerical (date format: yyyy-mm-dd).

Type: Continuous.

Scale: Interval.

Source: UCDP/PRIO Conflict Termination Dataset, Version 2-2015 (Kreutz 2010).

9.29 Contested states and peace agreements (confpagree) | (B)

This indicator registers whether the contested state has cut a peace agreement with the parent states during a specific calendar year. The (confpagree) takes one of the two following values:

0 – If there was no peace agreement between the contested state and the parent state.

1 – If there was a peace agreement between the contested state and the parent state.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: UCDP/PRIO Conflict Termination Dataset, Version 2-2015 (Kreutz 2010).

9.30 Date of peace agreements (confpagreedate) | (B)

This indicator captures the date when the peace agreement was reached. If there was no peace agreement, the field for the date is left blank. For contested state case-years for which there were no registered peace agreements, the fields are marked with NA.

Format: Numerical (date format: yyyy-mm-dd).

Type: Continuous.

Scale: Interval.

Source: UCDP/PRIO Conflict Termination Dataset, Version 2-2015 (Kreutz 2010).

10 PART IX External support indicators

Data on external support provided to contested states are secured primarily from the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011), which includes data for each party in conflict on the level of external supporters for the 1975-2009 period. I make some further updates and make some structural changes to the dataset in four important respects.

First, the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011) includes information on external support only for conflict periods; thus, I update the data on external support to contested states to also cover periods during which contested states were in peace with their respective parent states. Data on external support during the peace period are not straightforward to capture since external support during peace periods tends to be invisible. For example, it is difficult to capture instances if, in the absence of conflict, troops, weapons, or access to territory is provided, as such forms of external support become visible mostly when conflict erupts. To tackle this gap, I use additional literature and other sources, such as UN archival records. Knowing who were external supporters during conflict periods from the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011) has been particularly useful, as it has made the searching process for the peace periods easier - in that it is very likely (but not always) that external supporters during conflict may continue the support for actors they supported during conflict periods.

Second, the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011) does not always include all the cases of contested states. In such circumstances, I use other sources that range from reports, secondary data, as well as various historical media archives from Lexis Nexis and Factiva to update such information. For consistency purposes, I identify external support to contested states not covered by the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011), by following the latter’s data description and coding procedures provided in its codebook.

Third, I make some structural changes to suit the structure of my dataset, as well as to make the empirical data available in the UCDP External Support – Disaggregated Level Dataset, Version 1.0 (Högbladh, Pettersson, and Themnér 2011) more useful for an array of research needs. While I maintain each form of external support, such as troops, weapons, access to territory, etc., at an indicator level (each of them as separate indicators), I make a further distinction on the type of external support offered for each form of support. This means that I also identify whether each form of external support is provided by a state, non-state, neighboring, or Great Power actor.

The last structural change I make to the UCDP External Support – Disaggregated Level Dataset, Version 1 (Högbladh, Pettersson, and Themnér 2011) is that I include such support as logistical, training, intelligence, into a single indicator called “other” depending on the type of actor providing such support. This is mainly done because it is particularly difficult to dissect such forms of support separately for each contested state, for each of their years of existence, and each of their conflict and peace periods.

10.1 External support (exsupp) | (B)

This indicator registers whether contested states have received any form of external support from both state and non-state actors in each observed year. The (exsupp) takes one of the following two values:

1 – If the contested state has received any form of external support.

0 – If the contested state has received no external support at all.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.2 Form of external support: troops (extroops) | (B)

This indicator registers whether contested states have received external support in the form of troops by an external actor in each observed year. The (extroops) does not capture cases when troops are provided to or are present in the contested state as internationally commissioned peacekeeping troops (usually) with the agreement of the parent state. For example, the fact that India sent peacekeeping troops to Tamil Eelam as part of an agreement with Sri Lanka is not registered here. The presence of such peacekeeping troops is registered as separate indicators in another section. Turkish forces in Northern Cyprus, or Armenian forces in Nagorno-Karabakh, for instance, are captured by this indicator. The (extroops) takes one of the following values:

0 – If the contested state has not received support in the form of troops by outside actors.

1 – If the contested state has received support in the form of troops by outside non-state actors.

2 – If the contested state has received support in the form of troops by outside state actors.

3 – If the contested state has received support in the form of troops by outside state and non-state actors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source:UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.3 Form of external support: weapons (exweapons) | (B)

This indicator registers whether contested states have received external support in the form of weapons by an external actor in each observed year. The (exweapons) takes one of the following values:

0 – If the contested state has not received support in the form of weapons by outside actors.

1 – If the contested state has received support in the form of weapons by outside non-state actors.

2 – If the contested state has received support in the form of weapons by outside state actors.

3 – If the contested state has received support in the form of weapons by outside state and non-state actors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.4 Form of external support: access to territory (exaccess) | (B)

This indicator registers whether contested states have received external support in the form of access to territory by an external actor in each observed year. Access to territory usually is understood as the provision of a safe haven for rebels fighting a central government at times of conflict (in this case, contested states fighting their parent states). Additionally, access to territory here stretches beyond the provision of a safe havens during times of conflict to access to territory in times of peace. For example, the ability of government representatives or leadership of contested states and their populations to be able to visit and organize meetings in another (usually) neighboring state. Although cases like Abkhazia, South Ossetia, Kosovo, Northern Cyprus and some others which have mostly been living in the absence of conflict are registered to have had access to the territory of other (usually) neighboring states. The (exaccess) takes one of the following values:

0 – If the contested state has not been granted access to a territory by outside actors.

1 – If the contested state has been granted access to a territory by outside non-state actors.

2 – If the contested state has been granted access to a territory by outside state actors.

3 – If the contested state has been granted access to a territory by outside state and non-state actors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.5 Form of external support: financial/economic (exfinance) | (B)

This indicator registers whether contested states have received external support in the form of finance or economic boost by an external actor in each observed year. Financial and/or economic support is understood any provision of financial or economic aid, which does not include loans that need to be paid back; diaspora support; and other one-directional financial aid to contested states. The (exfinance) takes one of the following values:

0 – If the contested state has not received financial/economic support by outside actors.

1 – If the contested state has received financial/economic support by outside non-state actors.

2 – If the contested state has received financial/economic support by outside state actors.

3 – If the contested state has received financial/economic support by outside state and non-state actors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.6 Form of external support: other (exother) | (B)

This indicator registers whether contested states have received other forms of external support by an external actor in each observed year. Other forms of external support include either one or all other forms of external support that UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) indicate, which include: (i) “Access to military or intelligence infrastructure”; (ii) “Materiel/Logistics”; (iii) “Training/Expertise”; (iv) “Intelligence material”; and (v) “Other forms of support” The (exother) includes these other forms of external support in a single indicator for practical data related reasons. It is often difficult to capture such (other) forms of external support separately for each contested state for each of the years. Nonetheless, gathering them in a single indicator can be useful to record various other forms of support that contested states may have received, and to distinguish between those that have not received any form of external support at all. The (exother) takes one of the following values:

0 – If the contested state has not received other forms of support by outside actors.

1 – If the contested state has received other forms of support by outside non-state actors.

2 – If the contested state has received other forms of support by outside state actors.

3 – If the contested state has received other forms of support by outside state and non-state actors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.7 Names of external supporters (exname) | (B)

This indicator registers the names of actors, both state and non-state, who provided any form of external support to contested states. The (exname) offers a numbered list of actors, the order of which does not matter. For example, the (exname) for Biafra registers the following numbered list of actors who provided support to Biafra during certain years of the latter’s existence: (1)France; (2)Portugal; (3)China; (4)Tanzania, (5)Gabon, (6)Ivory Coast; (7)Zambia. Moreover, the (exname) offers a way to identify when external support has been provided to different parties inside the contested state through the (→) sign, which marks the direction towards which actor inside the contested state was the external support provided. For example, the (exname) for Eritrea in 1977 is: (1)Sudan→(ELF/EPLF); (2)Kuwait→(ELF/EPLF); (3)Saudi Arabia→(ELF); (4)Iraq→(ELF); (5)Diaspora→(EPLF); (6)UEA→(EPLF), noting the different parties inside the contested state that received support from different external supporters.

Format: Textual

Type: n/a

Scale: n/a

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.8 External support by neighbors (exneigbhour) | (B)

This indicator registers whether contested states have received any form of support from any neighbor in each observed year. The neighbor is defined here as a state or non-state actor, which either shares a land border with the contested state or is located within 100 kilometers from the border of the contested state when non-strict contiguity between pairs is present due to possible spans of water separating actors. This is done following Gleditsch and Ward (2016) suggestion of establishing a minimum-distance to measure proximity between two pairs of actors, as well as Idean Salehyan’s (2007) proposition of establishing 100 kilometers between two countries - a distance which seems to be suitable when studying the interaction between non-state actors and their neighboring state supporters. The (exneigbhour) takes one of the following two values:

1 – If the contested state has received any form of neighbor support.

0 – If the contested state has received no neighbor support at all.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.9 Neighbor support with troops (extroopsN) | (B)

This indicator registers whether contested states have received external support in the form of troops from a neighbor in each observed year. The (extroopsN) does not capture cases when troops are provided to or are present in the contested state as internationally commissioned peacekeeping troops (usually) with the agreement of the parent state. For example, the fact that Russia sent peacekeeping troops to Abkhazia as part of an agreement with Georgia between the early 1990s up to 2008 is not registered here. The presence of such peacekeeping troops is registered as separate indicators in another section. Turkish forces in Northern Cyprus, or Armenian forces in Nagorno-Karabakh, for instance, are captured by this indicator. The (extroopsN) takes one of the following values:

0 – If the contested state has not received support in the form of troops by their neighbors.

1 – If the contested state has received support in the form of troops by their non-state neighbors.

2 – If the contested state has received support in the form of troops by their state neighbors.

3 – If the contested state has received support in the form of troops by their state and non-state neighbors.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source:UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.10 Neighbor support with weapons (exweaponsN) | (B)

This indicator registers whether contested states have received external support in the form of weapons from a neighbor in each observed year. The (exweaponsN) takes one of the following values:

0 – If the contested state has not received support in the form of weapons by their neighbors.

1 – If the contested state has received support in the form of weapons by their non-state neighbors.

2 – If the contested state has received support in the form of weapons by their state neighbors.

3 – If the contested state has received support in the form of weapons by their state and non-state neighbors.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.11 Neighbor support with access to territory (exaccessN) | (B)

This indicator registers whether contested states have received external support in the form of access to a territory from a neighbor in each observed year. Additionally, access to territory here stretches beyond the provision of a safe havens during times of conflict to access to territory in times of peace. For example, the ability of government representatives or leadership of contested states and their populations to be able to visit and organize meetings in another (usually) neighboring state. Although cases like Abkhazia, South Ossetia, Kosovo, Northern Cyprus, and some others which have mostly been living in the absence of conflict are registered to have had access to the territory of other (usually) neighboring states. The (exaccessN) takes one of the following values:

0 – If the contested state has not been granted access to territory by their neighbors.

1 – If the contested state has been granted access to territory by their non-state neighbors.

2 – If the contested state has been granted access to territory by their state neighbors.

3 – If the contested state has been granted access to territory by their state and non-state neighbors.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.12 Neighbor support with financial/economic boost (exfinanceN) | (B)

This indicator registers whether contested states have received external support in the form of finance or economic boost by a neighbor in each observed year. Financial and/or economic support is understood as any provision of financial or economic aid, which does not include loans that need to be paid back; diaspora support; and other one-directional financial aid to contested states. The (exfinanceN) takes one of the following values:

0 – If the contested state has not received financial/economic support by neighbors.

1 – If the contested state has received financial/economic support by non-state neighbors.

2 – If the contested state has received financial/economic support by state neighbors.

3 – If the contested state has received financial/economic support by state and non-state neighbors.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.13 Other forms of support by neighbors (exotherN) | (B)

This indicator registers whether contested states have received other forms of external support from any neighbor in each observed year. Other forms of external support by neighbors include either one of all other forms of external support that UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) indicate, which include: (i) “Access to military or intelligence infrastructure”; (ii) “Materiel/Logistics”; (iii) “Training/Expertise”; (iv) “Intelligence material”; and (v) “Other forms of support”. The (exotherN) includes these other forms of external support in a single indicator for practical data related reasons. It is often difficult to capture such (other) forms of external support separately for each contested state for each of the years. Nonetheless, gathering them in a single indicator can be useful to record various other forms of support that contested states may have received, and to distinguish between those that have not recieved any form of external support at all. The (exotherN) takes one of the following values:

0 – If the contested state has not received other forms of external support by their neighbors.

1 – If the contested state has received other forms of external support by their non-state neighbors.

2 – If the contested state has received other forms of external support by their state neighbors.

3 – If the contested state has received other forms of external support by their state and non-state neighbors.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.14 Names of neighbor supporters (exNname) | (B)

This indicator registers the names of neighboring actors who provided any form of neighbor support to contested states. The (exNname) offers a numbered list of actors, the order of which does not matter. The (exNname) offers a way to identify when neighbor support has been provided to different parties inside the contested state through the (→) sign, which marks the direction towards which actor inside the contested state was the external support provided.

Format: Textual

Type: n/a

Scale: n/a

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.15 External support by Great Powers (exunsc) | (B)

This indicator registers whether contested states have received any form of support from a Great Power in each observed year. I adopt Bridget Coggin’s (2011) definition of Great Powers, which includes the five Permanent Members of the UN Security Council, namely, the U.S., the U.K., France, China, and Russia, as well as Japan and Germany after 1990. The (exunsc) takes one of the following two values:

1 – If the contested state has received any form of Great Power support.

0 – If the contested state has received no Great Power support at all.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.16 Great Power support with troops (extroopsGP) | (B)

This indicator registers whether contested states have received external support in the form of troops by a Great Power in each observed year. The (extroopsGP) does not capture cases when troops are provided to or are present in the contested state as internationally commissioned peacekeeping troops (usually) with the agreement of the parent state. For example, the fact that the U.S., the U.K, and France, among others, sent peacekeeping troops to Kosovo as part of an agreement with Serbia since 1999, is not registered here. The presence of such peacekeeping troops is registered as separate indicators in another section. The (extroopsGP) takes one of the following values:

0 – If the contested state has not received support in the form of troops by Great Powers.

1 – If the contested state has received support in the form of troops by non-state Great Powers

2 – If the contested state has received support in the form of troops by state Great Powers

3 – If the contested state has received support in the form of troops by state and non-state Great Powers.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source:UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.17 Great Power support with weapons (exweaponsGP)

This indicator registers whether contested states have received external support in the form of weapons by Great Powers in each observed year. The (exweaponsGP) takes one of the following values:

0 – If the contested state has not received support in the form of weapons by Great Powers.

1 – If the contested state has received support in the form of weapons by non-state Great Powers

2 – If the contested state has received support in the form of weapons by state Great Powers

3 – If the contested state has received support in the form of weapons by state and non-state Great Powers.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.18 Great Power support with access to territory (exaccessGP) | (B)

This indicator registers whether contested states have received external support in the form of access to a territory by Great Powers in each observed year. Additionally, access to territory here stretches beyond the provision of a safe havens during times of conflict to access to territory in times of peace. For example, the ability of government representatives or leadership of contested states and their populations to be able to visit and organize meetings in another (usually) neighboring Great Power state. Although cases like Abkhazia, South Ossetia, Lugansk Republic, and some others which have mostly been living in the absence of conflict, are registered to have had access to the territory of other (usually) neighboring Great Power states. The (exaccessGP) takes one of the following values:

0 – If the contested state has not been granted access to territory by Great Powers.

1 – If the contested state has been granted access to territory by non-state Great Powers

2 – If the contested state has been granted access to territory by state Great Powers

3 – If the contested state has been granted access to territory by state and non-state Great Powers.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.19 Great Power support with financial/economic boost (exfinanceGP) | (B)

This indicator registers whether contested states have received external support in the form of finance or economic boost by a Great Power in each observed year. Financial and/or economic support is understood any provision of financial or economic aid, which does not include loans that need to be paid back; diaspora support; and other one-directional financial aid to contested states. The (exfinanceGP) takes one of the following values:

0 – If the contested state has not received financial/economic support by Great Powers.

1 – If the contested state has received financial/economic support by non-state Great Powers.

2 – If the contested state has received financial/economic support by state Great Powers.

3 – If the contested state has received financial/economic support by state and non-state Great Powers.

Format: Numerical.

Type: Categorical.

Scale: Nominal

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.20 Other forms of support by Great Powers (exotherGP) | (B)

This indicator registers whether contested states have received other forms of external support by Great Powers in each observed year. Other forms of external support by Great Powers include either one of all other forms of external support that UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011) indicate, which include: (i) “Access to military or intelligence infrastructure”; (ii) “Materiel/Logistics”; (iii) “Training/Expertise”; (iv) “Intelligence material”; and (v) “Other forms of support”. These other forms of external support also include Diaspora support. The (exotherGP) includes these other forms of external support in a single indicator for practical data related reasons. It is often difficult to capture such (other) forms of external support separately for each contested state for each of the years. Nonetheless, gathering them in a single indicator can be useful to record various other forms of support that contested states may have recieved, and to distinguish between those that have not recieved any form of external support at all. The (exotherGP) takes one of the following values:

0 – If the contested state has not received other forms of support by Great Powers.

1 – If the contested state has received other forms of support by non-state Great Powers

2 – If the contested state has received other forms of support by state Great Powers

3 – If the contested state has received other forms of support by state and non-state Great Powers.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

10.21 Names of Great Power supporters (exGPname) | (B)

This indicator registers the names of Great Power actors who provided any form of external support to contested states. The (exGPname) offers a numbered list of actors, the order of which does not matter. The (exGPname) offers a way to identify when Great Power support has been provided to different parties inside the contested state through the (→) sign, which marks the direction towards which actor inside the contested state was the external support provided.

Format: Textual

Type: n/a

Scale: n/a

Source: UCDP External Support - Primary Warring Party Dataset (Högbladh, Pettersson, and Themnér 2011)

11 PART X: Foreign presence indicators

11.1 UN peacekeeping presence (unpk) | (A)

This indicator registers whether there has been a presence of a UN peacekeeping mission in either the contested state or the parent state. Peacekeeping presence here is defined according to the UN’s categorization, which means that it can take two forms: (i) peacekeeping operations and (ii) political missions (United Nations 2018e). Peacekeeping operations can also be those in which military force or weapons are not necessarily used, such as the UNOMIG operation deployed in Georgia in 1994, where the UNOMIG mission had the goal to observe the ceasefire agreement and conduct other monitoring tasks. Peacekeeping operations can also be fully-fledged military operations such as the ONUC in the D.R. Congo deployed in 1960, where UN peacekeepers were at some point tasked also to use force against contested states emerging out of D.R. Congo, such as Katanga and South Kasai. They can also be large civilian missions that administer another state or territory, such as the UNMIK mission deployed in Kosovo in 1999. The (unpk) indicator takes one of the following two values:

1 – If there is a UN peacekeeping presence in either the contested state or the parent state (or in both).

0 – If there is no UN peacekeeping presence in either the contested state or the parent state (or in both)

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

11.2 Number of UN peacekeeping personnel (unpknr) | (A)

This indicator registers the total number of personnel deployed as part of a UN peacekeeping presence in either the contested state or the parent state. For cases where there is no peacekeeping presence, this indicator takes the value of 0.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: To be provided upon request

11.3 UN peacekeeping name (unpkname) | (A)

This indicator registers the name of the UN peacekeeping mission present in either the contested state or the parent state. For cases where there is no peacekeeping presence, this indicator takes the value of NA.

Format: Textual

Type: Categorical.

Scale: Nominal.

Source: To be provided upon request

11.4 UN peacekeeping presence (unpkpresence) | (A)

This indicator registers the location of the UN peacekeeping mission operations. The (unpkpresence) can be helpful in understanding also the mandate of a particular UN peacekeeping presence, such as, whether they are mandated to stay and protect the contested state, or to re-introduce the authority of the parent state over the contested state, and so on. The (unpkpresence) takes one of the following four values:

0 – If there is no UN peacekeeping presence in either the contested state or the parent state (or in both).

1 – If the UN peacekeeping presence is in both, the parent state and the contested state. These also include those missions that stand at the border between the contested state and the parent state, such as the UNOMIG in Georgia, or the UNFICYP in Cyprus, which divide the belligerents and observe the ceasefire lines between the two parties.

2 – If the UN peacekeeping presence is in the parent state only. These include such missions as the UNPROFOR in Bosnia or the ONUC in the Republic of Congo. Note: for analytical reasons, the UN peacekeeping missions, which were located in the respective contested states but mandated to re-introduce the respective parent state’s authority over the contested state, are also part of this category.

3 – If the UN peacekeeping presence is in the contested state only. These include such missions as the UNMIK in Kosovo or the MINURSO in Western Sahara.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: To be provided upon request

11.5 Non-UN peacekeeping presence (nonunpk) | (A)

This indicator registers whether there is a non-UN peacekeeping mission in either the contested state or the parent state. Non-UN peacekeeping mission here refers to a foreign mission multilaterally approved with the consent of the parent state, which is present either in the parent state or the contested state. For instance, the CIS peacekeeping force in Georgia between 1993 to 2008; KFOR peacekeeping force in Kosovo since 1999; or India’s peacekeeping force in Tamil Eelam between 1987 to 1990 are part of the non-UN peacekeeping presence. This means that the presence of these missions can be UN approved or accepted, but the UN does not run them. The (nonunpk) does not include the presence of foreign troops/missions without the consent of the parent state, such as the Armenian troops in Nagorno-Karabakh, or Belgian troops in Katanga, and other cases alike. Such missions are registered as foreign troops as separate indicators. The (nonunpk) takes one of the following two values:

1 – If there is a non-UN peacekeeping presence in either the contested state or the parent state (or both).

0 – If there is no non-UN peacekeeping presence in either the contested state or the parent state (or both).

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

11.6 Number of non-UN peacekeeping personnel (nonunpknr) | (A)

This indicator registers the total number of personnel deployed as part of a non-UN peacekeeping presence in either the contested state or the parent state. For cases where there is no non-UN peacekeeping presence, this indicator takes the value of 0.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: To be provided upon request

11.7 Non-UN peacekeeping name (nonunpkname) | (A)

This indicator registers the name of the non-UN peacekeeping mission present in either the contested state or the parent state. For cases where there is no peacekeeping presence, this indicator takes the value of NA.

Format: Textual

Type: Categorical.

Scale: Nominal.

Source: To be provided upon request

11.8 Non-UN peacekeeping presence (nonUNPKpresence) | (A)

This indicator registers the location of the non-UN peacekeeping mission operations. The (nonUNPKpresence) can be helpful in understanding also the mandate of a particular non-UN peacekeeping presence, such as, whether they are mandated to stay and protect the contested state, or to re-introduce the authority of the parent state over the contested state, and so on. The (nonUNPKpresence) takes one of the following four values:

0 – If there is no non-UN peacekeeping presence in either the contested state or the parent state (or in both).

1 – If the non-UN peacekeeping presence is in both, the parent state and the contested state.

2 – If the non-UN peacekeeping presence is in the parent state only.

3 – If the non-UN peacekeeping presence is in the contested state only.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: To be provided upon request

11.9 Other missions in the contested state (othmission) | (A)

This indicator registers whether there are other missions in the contested state (not in the parent state). Other missions are not UN peacekeeping missions and not non-UN peacekeeping missions. Typically, they are regional missions that are established inside the contested states. Some examples include the EU Rule of Law Mission in Kosovo (EULEX), or the Organization of African Union Mission in Comoros (established in Anjouan). The othmission) registers the number of such missions, and it takes the following values:

0 – If there is no other peacekeeping mission in the contested state.

1 – If there is one other peacekeeping mission in the contested state.

2 – If there are two other peacekeeping missions in the contested state.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: To be provided upon request

11.10 Number of other missions’ personnel (othmissionnr) | (A)

This indicator registers the total number of personnel deployed as part of the other missions in contested states – typically a range. The (othmissionnr) ranges from less than (<); to more than (>); or the range between two numbers (-) as approximate specifiers.

Format: Numerical (with range specifiers).

Type: Continuous.

Scale: Ratio.

Source: To be provided upon request

11.11 Other missions’ name(s) (othmissionname) | (A)

This indicator registers the name of the other missions present in the contested state. For cases where there is no peacekeeping presence, this indicator takes the value of NA.

Format: Textual

Type: n/a.

Scale: n/a.

Source: To be provided upon request

11.12 Foreign troops in the contested state (foreigntroops) | (A)

This indicator registers whether there has been a presence of foreign troops in the contested state, which are not part of the above missions. These troops are usually present in the contested state outside any multilateral agreement and without the consent of the parent state. The (foreigntroops) represents the troops that are more permanently based in the contested state, meaning they are not only temporarily provided during times of war; they are also stationed in times of peace. For instance, the (foreigntroops) captures Russian troops in Abkhazia and South Ossetia after 2008; Armenian troops in Nagorno-Karabakh since its proclamation of independence in the early 1990s; Indian troops in Bangladesh during the 1971-1972 period, and alike. The (foreigntroops) takes one of the following two values:

1 – If there are foreign troops stationed in the contested state.

0 – If there are no foreign troops stationed in the contested state.

Format: Numerical.

Type: Categorical.

Scale: Nominal/dichotomous.

Source: To be provided upon request

11.13 Number of foreign troops in the contested state (foreigntroopsnr) | (A)

This indicator registers the total number of foreign troops stationed in the contested state. For cases where there are no foreign troops in the contested state, the (foreigntroopsnr) takes the value of 0.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: To be provided upon request

11.14 Foreign troops names (foreigntroopsname) | (A)

This indicator registers the name of the party providing foreign troops to the contested state. For cases where there are no foreign troops stationed, this indicator takes the value of NA.

This indicator captures the number of international governmental organizations (IGOs) that the contested state has been a partial member of. Being a partial member or not a full member of an IGO means that the contested state does not enjoy all the rights as other (UN member) states enjoy, which are part of the same IGO. The (nonfulligo); thus, captures such cases as when contested states become members of IGOs with an “observer” status, “associate” membership, “non-voting” membership, and alike.

Format: Numerical.

This indicator captures the number of times the contested state has featured in the International Court of Justice (ICJ) press releases.

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: ICJ (2018)

14 PART XIII: Other indicators

14.1 Regions: alternative 1 (region1) | (C)

This indicator captures the regions where contested states belong. There are many ways in which regions can be defined and categorized, but (region1) categorizes regions where contested states belong according to the general UCDP/PRIO regional categorizations, as follows:

1 – Europe.

2 – Middle East.

3 – Asia.

4 – Africa.

5 – Americas.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: Reference to general UCDP/PRIO datasets and standards of categorization

14.2 Regions: Alternative 2 (region2) | (C)

This indicator captures the regions where contested states belong. There are many ways in which regions can be defined and categorized, but (region2) categorizes regions where contested states belong according to Florea’s (2014) regional categorizations, as follows:

1 – Western Europe.

2 – Central and Eastern Europe.

3 – Eurasia (post-Soviet East European countries and the Caucasus).

4 – Middle East and North Africa (MENA).

5 – Sub-Saharan Africa.

6 – Central Asia.

7 – Southeast Asia.

8 – Americas.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: Reference to general UCDP/PRIO datasets and standards of categorization

14.3 Contested state decades (decade) | (A)

This indicator captures the decade to which each contested state belonged during each observed year.

Format: Numerical.

Type: Categorical.

Scale: Nominal.

Source: Author’s observation from (yearcs)

14.4 Number of contested states (nrcs) | (A)

This captures the number of contested states during each year of the existence of all contested states.

Format: Numerical.

Type: Continuous.

Scale: Interval.

Source: Author’s observation from (outcome)

15 PART XIV: Transformation of indicators

This section includes the categorization, normalization, or rescaling of some of the existing indicators. Decisions about the number of categories or the rescaling levels are taken based on their utility when used and/or combined with other existing (usually categorical) indicators in this dataset to generate higher-order conceptual constructions and meanings. The use of these indicators is optional, and research can use the existing indicators to categorize and rescale them according to their research needs.

15.1 Number of factions in contested states categorized (factions_4_cat) | (A)

This is a categorized indicator of (factions) based on 3 quantiles of the mean values of each case over their life course. The (factions_4_cat) takes one of the below values depending on the quantiles of the number of factions that exist in each contested state:

1 – Very high level of factionalism (higher than the 75^th quantile, capturing more than 4 factions a contested state).

2 – High level of factionalism (between the 50^th and 75^th quantile, capturing between 3 and 4 factions in a contested state).

3 – Low level of factionalism (between the 25^th and 50^th quantile, capturing between 2 and up to 3 factions in a contested state).

4 – Very low level or no factionalism in contested states (lower than the 25^th quantile, capturing up to 2 factions in a contested state).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (factions)

15.2 Number of factions in contested states normalized and rescaled (factions_r_log) | (A)

This is the log of (factions) rescaled to a continuous indicator taking values between 1 (very high factionalism in contested states) to 4 (very low or no factionalism in contested states).

Format: Numerical.

Type: Continuous.

Scale: Interval.

Source: Log normalization and rescaling of (factions)

15.3 Ratio of area under the control of contested states categorized (areacontrolcs_4_cat) | (A)

This is a categorized indicator of (areacontrolcs) based on 3 quantiles of the mean values of each case over their life course. The (areacontrolcs_4_cat) takes one of the below values depending on the quantiles of each contested state’s area of territorial control:

1 – Low level of territorial control (lower than the 25^th quantile, capturing up to 53 percent control over the territory).

2 – Moderate to a high level of territorial control (between the 25^th and 50^th quantile, capturing between 53 and 91 percent control over the territory).

3 – Very high level of territorial control (between the 50^th and 75^th quantile, capturing between 91 and 100 percent control over the territory).

4 – Extremely high level of territorial control (higher than the 75^th quantile, capturing more than 100 percent control over the territory).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (areacontrolcs)

15.4 Ratio of area under the control of contested states normalized and rescaled (areacontrolcs_r_log) | (A)

This is the log of (areacontrolcs) rescaled to a continuous indicator taking values between 1 (low level of territorial control f) to 4 (extremely high level of territorial control).

Format: Numerical.

Type: Continuous.

Scale: Interval.

Source: Log normalization and rescaling of (areacontrolcs)

15.5 Parent states’ days in conflict excluding contested state categorized (confpdays_6_cat) | (A)

This is a categorized indicator of (confpdayswoCS) based on 5 quantiles of the mean values of each case over their life course. The (confpdays_6_cat) takes one of the below values depending on the quantiles of each parent state’s days at war other than those with the contested state:

0 – Not in conflict during a year (lower than or equal to the 10^th quantile, capturing up to 0 days in conflict).

1 – Not much in conflict during a year (between 10^th and 30^th quantile, capturing around 0.6 days in conflict).

2 – Somewhat in conflict during a year (between the 30^th and 50^th quantile, capturing between 0.7 and 37 days in conflict).

3 – Partially in conflict during a year (between the 50^th and 70^th quantile, capturing between 38 and 113 days in conflict).

4 – Considerably in conflict during a year (between the 70^th and 90^th quantile, capturing between 114 and 336 days in conflict).

5 – Fully in conflict during a year (lower than or equal to the 90^th quantile, capturing more than 337 days in conflict).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (confpdayswoCS)

15.6 Parent states’ days in conflict excluding contested state normalized and rescaled (confpdays_r_croot) | (A)

This is the cube root of (confpdayswoCS) rescaled to a continuous indicator taking values between 0 (no conflict during any given year) to 5 (Fully in conflict during any given year).

Format: Numerical.

Type: Continuous.

Scale: Ratio

Source: Log normalization and rescaling of (confpdayswoCS)

15.7 Percent of population mobilized in parent states categorized (troopspcap_6_cat) | (A)

This is a categorized indicator of (troopspcap) based on 5 quantiles of the mean values of each case over their life course. The (troopspcap_6_cat) takes one of the below values depending on the quantiles of each parent state’s ability to mobilize their population:

0 – Very high mobilization capacity (higher than the 90^th quantile, capturing more than 2.20 percent of the mobilized population).

1 – High mobilization capacity (between 70^th and 90^th quantile, capturing the range between 0.69 and 2.20 percent of the mobilized population).

2 – Moderate to high mobilization capacity (between the 50^th and 70^th quantile, capturing the range between 0.46 and 0.68 percent of the mobilized populatio).

3 – Moderate to low mobilization capacity (between the 30^th and 50^th quantile, capturing the range between 0.28 and 0.45 percent of the mobilized population).

4 – Low mobilization capacity (between the 10^th and 30^th quantile, capturing the range between 0.09 and 0.27 percent of the mobilized population).

5 – Very low mobilization capacity (lower than or equal to the 10th quantile, capturing up to 0.08 percent of the mobilized population).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (troopspcap)

15.8 Percent of population mobilized in parent states normalized and rescaled (troopspcap_r_croot) | (A)

This is the cube root of (troopspcap) rescaled to a continuous indicator taking values between 0 (very high mobilization capacity) to 5 (very low mobilization capacity).

Format: Numerical.

Type: Continuous.

Scale: Ratio

Source: Log normalization and rescaling of (troopspcap)

15.9 Per capita military expenditure of parent states categorized (milexcap_6_cat) | (A)

This is a categorized indicator of (milexcap) based on 5 quantiles of the mean values of each case over their life course. The (milexcap_6_cat) takes one of the below values depending on the quantiles of each parent state’s military expenditure per capita:

0 – Very high military expenditure (higher than the 90^th quantile, capturing more than 610 USD per capita expenditure).

1 – High military expenditure (between 70^th and 90^th quantile, capturing between 85 and 609 USD per capita expenditure).

2 – Moderate to military expenditure (between the 50^th and 70^th quantile, capturing between 75 and 84 USD per capita expenditure).

3 – Moderate to low military expenditure (between the 30^th and 50^th quantile, capturing between 26 and 74 USD per capita expenditure).

4 – Low military expenditure (between the 10^th and 30^th quantile, capturing between 9 and 25 USD per capita expenditure).

5 – Very low military expenditure (lower than or equal to the 10^th quantile, capturing up to 8 USD per capita expenditure).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (milexcap)

15.10 Per capita military expenditure of parent states normalized and rescaled (milexcap_r_croot) | (A)

This is the cube root of (milexcap) rescaled to a continuous indicator taking values between 0 (very high military expenditure) to 5 (very low military expenditure).

Format: Numerical.

Type: Continuous.

Scale: Ratio

Source: Log normalization and rescaling of (milexcap)

15.11 Parent state’s GDP per capita categorized (cgdppc_6_cat) | (A)

This is a categorized indicator of (cgdppc) based on 5 quantiles of the mean values of each case over their life course. The (cgdppc_6_cat) takes one of the below values depending on the quantiles of each parent state’s GDPper capita:

0 – Very high GDP per capita (higher than the 90^th quantile, capturing more than 13,355 USD per capita GDP).

1 – High GDP per capita (between 70^th and 90^th quantile, capturing between 6,728 and 13,355 USD per capita GDP).

2 – Moderate to high GDP per capita (between the 50^th and 70^th quantile, capturing between 3,546 and 6,727 USD per capita GDP).

3 – Moderate to low GDP per capita (between the 30^th and 50^th quantile, capturing between 1,735 and 3,545 USD per capita GDP).

4 – Low GDP per capita (between the 10^th and 30^th quantile, capturing between 1,161-1,1734 USD per capita GDP).

5 – Very low GDP per capita (lower than or equal to the 10^th quantile, capturing up to 1,160 USD per capita GDP).

Format: Numerical.

Type: Categorical.

Scale: Ordinal.

Source: Categorization of (cgdppc)

15.12 Parent state’s GDP per capita normalized and rescaled (cgdppc_r_croot) | (A)

This is the log of (cgdppc) rescaled to a continuous indicator taking values between 0 (very high GDP per capita) to 5 (very low GDP per capita).

Format: Numerical.

Type: Continuous.

Scale: Ratio.

Source: Log normalization and rescaling of (cgdppc)

16 PART XV: Indicators of a higher order of abstraction (aggregation)

This section includes aggregated indicators generated from the existing indicators to capture the conceptual phenomena of a higher order of abstraction.

16.1 External support score (extsupport)

There is no generally agreeable way how to model the measurement for external support to contested states. Nevertheless, ample ideas in doing so can be utilized from the literature, which focuses on external support to non-state actors. The first specification that can be made is with regard to the type of actors providing external support. With one recent exception (see: Sawyer, Cunningham, and Reed 2017), the overwhelming majority of authors analyzing external support to separatists, rebels, and insurgencies focus on state actors as suppliers of external support (Saideman 2001, @salehyanTransnationalRebelsNeighboring2007, @salehyanNoShelterHere2008, @schultzEnforcementProblemCoercive2010, @coggins_friends_2011, @carterBlessingCurseState2012, @florea_facto_2017). External support by non-state actors has not found much place among actors that have been analyzed as external supporters.

Second, Salehyan (2007), Salehyan (2008), and Coggins (2011), among others, already have provided strong arguments in making a distinction among types of state actors providing external support. Salehyan (2007) suggests that the provision of a safe havens by neighboring states to weak rebels fighting much stronger central governments has a strong impact on rebels’ survival. Sometimes, when neighboring states are weak and unable to fully control their territories often open a favorable gap for rebels to seek shelter in neighboring states, where they reorganize and resupply to fight back their central governments (Salehyan 2008). When “rebels can transit neighboring states (either through the connivance of an allied government or because of its weakness), it becomes far harder to defeat them” (Byman et al. 2001, 86). Additionally, in their categorization of external support to rebels, Byman et al. (2001) put the provision of a safe havens among the critical forms of external support together with the provision of direct military aid through troops or financial resources. This is because providing rebels with access to territory enables them to secure other forms of support. Many examples abound. Saideman (2001) considers the provision of a safe havens to Katanga by the Federation of Rhodesia and Nyasaland (British Central Africa) as important as the Belgians’ support to the Katangese. One form of support that the Federation of Rhodesia and Nyasaland provided to Katanga, which Belgium as a non-neighboring state could not, is the ability for Katanga to move mercenaries through the Federation’s territory back to Katanga. Also, “[s]mugglers ran arms and supplies across the border, including the fighter aircraft that attacked UN troops during the first two rounds of UN action. Katangan minerals were transported across Rhodesia for sale elsewhere, providing Tshombe with the hard currency necessary for the purchase of more arms, equipment, and mercenaries [emphasis mine]” (Saideman 2001, 46).

In addition to neighboring states, the provision of support by Great Powers has been central to Coggin’s (2011) argument of why some secessionist movements not only survive but also manage to, in the end, get the international legitimation of their statehood. Great Power support to contested states matters because such actors occupy the “high places” in world politics, such as IGOs and other fora where decisions about and matters of peace and war are discussed (Coggins 2011). Great Powers hold veto powers for matters of peace and conflict, especially in the UNSC. IGOs also create “empirically identifiable social networks that help to define the conditions under which acts of aggression or cooperation can be rational strategies of action in international relations. It is our core contention that interstate military aggression is not simply a result of bargaining failure but is suppressed or encouraged by the relative positions states occupy in the larger network of IGOs, which promote common beliefs and alter the distribution of social power” (Hafner-Burton and Montgomery 2006, 4). According to Coggins (2011), the historical evidence suggests that should Great Powers provide support to secessionist movements, they are more likely to survive than to die. Thus, my next specification that I make is that I focus on the provision of external support by the neighboring states (in the form of access to the territory), and by Great Powers (any kind of support) as an effective means to secure the survival of contested states.

The third specification I make is with regards to the question that becomes evident: how to deal with different forms of external support and whether some forms of support, such as provision of troops or weapons, should be weighed more than other forms of support, such as the provision of financial/economic support and other forms. Saideman (2001) weighs forms of external support based on intensity (between 0 - no external support at all, to 4 - highly intense forms of external support), as well as based on breadth, that is the number of countries providing external support. So do Byman et al. (2001) provide three categories of forms of external support based on whether they are crucial, valuable, or minor forms of external support. Yet, both give indications that measuring the importance of different forms of external support is difficult. Byman et al. go on to critically suggest that “[m]easuring the impact of outside support is difficult, because its value is always relative”, and the impact of one or another form of support provided to rebels depends on the stage of conflict, what the rebels need and demand, as well as the impact that one form of support may have on another (2001, 104). For example, rebel groups that may possess enough weapons to fight the central government may be more in need of political support. Rebels that are at the initial stage of conflict may value troops and training more than they may value political support or weapons. Likewise, the provision of financial support may have an important impact on rebels, as they may divert these funds to obtain weapons by themselves, or it can be used for civilian needs and can make it unpredictable for central governments to judge the intensity of external support that rebels may be receiving (Sawyer, Cunningham, and Reed 2017). From this discussion, we can learn that receiving only one form of support may not necessarily have a lower impact on rebels’ success, as compared to those receiving more forms of external support. Receiving more forms of external support as opposed to only one form can be a reflection of the situation on the ground whereby rebels receiving fewer forms of external support may simply possess enough of, for example, material and logistical capabilities, weapons or troops, but may simply need just safe havens. Given the unsettled debate on the importance and impact of varying forms of external support on the livelihood of rebel groups and other non-state actors, I opt for the specification where each form of external support is weighed equally and measured equally regardless of the varying forms of support they may receive.

In sum, the theoretical relevance of types of actors providing external support remains relevant, while there seems to be no strong theoretical argument for weighing external support differently. Therefore, my measurement of the external support score (extsupport) is based on the following criteria: only external support provided by neighboring states in the form of (i) access to territory, or any form of external support provided by Great Powers in the form of (i) troops, or (ii) weapons, or (iii) access to territory, or (iv) economic/financial, or (v) other forms are coded as 1, while the lack of any such provision is coded as 0. Note that such a dichotomous strategy is pursued by many authors measuring and analyzing external support to secessionist, rebel, and insurgent groups (see: Salehyan, Gleditsch, and Cunningham 2011; Salehyan 2007; Coggins 2011; and Sawyer, Cunningham, and Reed 2017). The (extsupport) is derived by the following expression:

\[ \begin{aligned} extsupport = & \left\{\begin{matrix} \mathbf{1} & \text{if} & exaccessN' & = & 1 &\lor & extroopsGP' & = & 1 &\lor & exweaponsGP' & = & 1 &\lor & exaccessGP' & = & 1 &\lor\\ \mathbf{0} & \text{if} & exaccessN' & = & 0 & \wedge & extroopsGP' & = & 0 & \wedge & exweaponsGP' & = & 0 & \wedge & exaccessGP' & = & 0 & \wedge \end{matrix}\right.\\ & \left.\begin{matrix} exfinanceGP' & = & 1 & \lor & exotherGP' & = & 1 \\ exfinanceGP' & = & 0 & \wedge & exotherGP' & = & 0 \end{matrix}\right\} \end{aligned} \]

(∨) represents the logical operator OR

(∧) represents the logical operator AND

(’) the prime symbol means that the original values of the indicators have been recorded to capture state sponsors only. This means that all 2 (state actors) and 3 (state and non-state actors) values are recoded to 1, and all 1 (non-state actors) values are recoded to 0, to capture state sponsors only.

16.2 External support score - alternative 1 (extsupport_alt_1)

There may be some doubts as to whether the original measurement model for external support is the right one. For example, one may raise a question the original measurement’s specification of capturing access to territory as the only form of support provided by neighbors while capturing all other forms of external support by Great Powers. I offer another alternative, which includes all forms of external support provided by neighbors that are also provided by Great Powers. Therefore, my alternative measurement of external support score (extsupport_alt_1) is based on the following criteria: any form of external support such as (i) troops, or (ii) weapons, or (iii) access to territory, or (iv) economic/financial, or (v) other forms provided by either neighbors or Great Powers are coded as 1, while the lack of any of such forms of external support is coded as 0. I keep the dichotomous strategy in this alternative measurement as such a strategy is widely pursued by many authors measuring and analyzing external support to secessionist, rebel, and insurgent groups (see: Salehyan, Gleditsch, and Cunningham 2011; Salehyan 2007; Coggins 2011; and Sawyer, Cunningham, and Reed 2017). The (extsupport_alt_1) is derived by the following expression:

\[ \begin{aligned} extsupport_{-}alt_{-}1 = & \left\{\begin{matrix} \mathbf{1} & \text{if} & extroopsN' & = & 1 &\lor & exweaponsN' & = & 1 &\lor & exaccessN' & = & 1 &\lor & exfinanceN' & = & 1 &\lor & exotheN' & = & 1 \lor\\ \mathbf{0} & \text{if} & extroopsN' & = & 0 & \wedge & exweaponsN' & = & 0 & \wedge & exaccessN' & = & 0 & \wedge & exfinanceN' & = & 0 & \wedge & exotheN' & = & 0 \wedge\\ \end{matrix}\right.\\ & \left.\begin{matrix} extroopsGP' & = & 1 & \lor & exweaponsGP' & = & 1 &\lor & exaccessGP' & = & 1 &\lor & exfinanceGP' & = & 1 &\lor & exotheGP' & = & 1\\ extroopsGP' & = & 0 & \wedge & exweaponsGP' & = & 0 & \wedge & exaccessGP' & = & 0 & \wedge & exfinanceGP' & = & 0 & \wedge & exotheGP' & = & 0 \\ \end{matrix}\right\} \end{aligned} \]

(∨) represents the logical operator OR

(∧) represents the logical operator AND

16.3 External support score - alternative 2 (extsupport_alt_2)

Questions can be raised further whether theoretical relevance should be given to the type of actors providing external support (i.e., neighbors and Great Powers as the only actors specified), instead of focusing on any type of state actor. I offer and test another alternative, which includes all state actors, regardless of whether or not they are neighbors or Great Powers. Thus, my other alternative measurement of external support score (extsupport_alt_2) is based on the following criteria: any form of external support such as as (i) troops, or (ii) weapons, or (iii) access to territory, or (iv) economic/financial, or (v) other forms provided by any state actor are coded as 1, while the lack of any such forms of external support is coded as 0. I keep the dichotomous strategy in this alternative measurement as such a strategy is widely pursued by many authors measuring and analyzing external support to secessionist, rebel, and insurgent groups (see: Salehyan, Gleditsch, and Cunningham 2011; Salehyan 2007; Coggins 2011; and Sawyer, Cunningham, and Reed 2017). The (extsupport_alt_2) is derived by the following expression:

\[ \begin{aligned} extsupport_{-}alt_{-}2 = & \left\{\begin{matrix} \mathbf{1} & \text{if} & extroops' & \lor & exweapons' & \lor & exaccess' &\lor \\ \mathbf{0} & \text{if} & extroops' & \wedge & exweapons' & \wedge & exaccess' & \wedge \end{matrix}\right.\\ & \left.\begin{matrix} exfinance' & \lor & exother' & = & 1 \\ exfinance' & \wedge & exother' & = & 0 \end{matrix}\right\} \end{aligned} \]

(∨) represents the logical operator OR

(∧) represents the logical operator AND

16.4 External support score - alternative 3 (extsupport_alt_3)

The final alternative of measuring external support is constructed to respond to another possible doubt that may be raised with regards to the theoretical (ir)relevance of weighing the intensity of external support. Despite the unsettled debate on the importance and impact of varying forms of external support on the livelihood of rebel groups and other non-state actors, I opt for the option where each form of external support is weighed equally, yet the intensity - whether contested states receive only one or more forms of external support - remains theoretically important. Saideman (2001) weighs forms of external support differently based on intensity (between 0 - no external support at all, and 4 - an intense form of external support) and breadth (based on the number of countries providing external support). So do Byman et al. (2001) provide three categories of forms of external support based on whether they are crucial, valuable, or minor forms of external support. Yet, both give indications that measuring the importance of different forms of external support is difficult. Therefore, my third alternative of measuring external support is based on the premise that if, for example, a contested state recieves only (i) weapons, the intensity of external support is coded as 1. If a contested state receives (i) weapons, and (ii) troops, and (iii) access to territory, the intensity of external support is coded as 3. A score of 1 is added for each additional form of external support that is provided to contested states. This is akin to what Florea (2017) pursues as a strategy to measure external support to contested states. Accordingly, my third alternative of external support measurement (extsupport_alt_3) is created based on the following conditions: (i) only external support provided by state actors is included; (ii) all possible forms of external support are weighted equally (troops, weapons, access to territory, financial/economic, and other forms); and (iii) all available forms of external support to contested stats are added together to derive (extsupport_alt_3). The (extsupport_alt_3) is derived by the following expression:

\[ extsupport_{-}alt_{-}3 = extroops' + exweapons' + exaccess' + exfinance' + exother' \]

The (extsupport_alt_3) can take one of the following values:

0 – If there was no support provided to contested states by outside state actors.

1 – If the support provided to contested states by outside state actors was of low intensity.

2 – If the support provided to contested states by outside state actors was of moderate intensity.

3 – If the support provided to contested states by outside state actors was of high intensity.

4 – If the support provided to contested states by outside state actors was of very high intensity.

5 – If the support provided to contested states by outside state actors was of extremely high intensity.

16.5 State-making score (statemaking)

The explanatory condition of state-making efforts aims to capture the degrees to which contested states have managed to develop state structures from the Weberian conception; meaning, their ability to build state institutions and achieve high degrees of monopoly of violence while managing to radiate their administrative and policy-making functions from centralized structures reaching outwardly to penetrate the territories they claim (see: Weber 2004; Mann 1984). Michael Mann (1984) asserts that Weber’s conception of the state includes the institutional element - the number and the kind of institutions that are in place, and the functional element - what states do with those institutions, such as implementing policy and penetrating society over wide stretches of territory. In doing so, he specifies the definition of the state as i) a differentiated set of institutions and personnel embodying; ii) centrality in the sense that political relations radiate outwards from a center to cover; iii) a territorially-demarcated area, over which it exercises; iv) a monopoly of authoritative binding rule-making, backed up by a monopoly of the means of physical violence (Mann 1984, 188). Mann’s (1984) updated Weberian conception of the state opens the way to delineate two ways in which state strength and/or weakness can be conceived of. The first is through the state’s despotic power - the ability of the state to remain autonomous from “civil-society” and take decisions without the need to negotiate with independent civil groups; the second, is the state’s infrastructural power - the capacity of the state to penetrate the society and implement political and logistical decisions) (Mann 1984). Thus, institutions, centrality, and territoriality remain key components of this conception

There are others who offer other ways to categorize and conceive of the dimensions of state strength and weakness. For example, adopting Theda Skocpol’s (1985) notion of varieties of state capacities, Hanson and Sigman (2013) focus on three core dimensions of state power, which include: i) the state’s extractive capacity - the state’s ability to extract revenues; ii) the state’s coercive capacity - the state’s ability to assert its monopoly on use of violence over a demarcated territory; and ii) the state’s administrative capacity - the ability of the state to implement policy and deliver goods and services. Desch (1996) speaks of the scope - the state’s reach over territory and population to mobilize people or provide services to them, and the cohesion - the state’s division or unity, namely the centrality of the state. Also, Grävingholt, Ziaja, and Kreibaum (2015) provide arguments for viewing state capacity from the perspective of three dimensions, namely: i) state’s authority - maintaining monopoly on violence and internal order; ii) capacity - ability to provide public goods and services; and iii) legitimacy - social acceptance of state’s presence without the necessity to use violence. Despite divergences, central to many of the above assertions remains the state’s institutions, centrality, and territoriality.

From this discussion, it becomes evident that an ideal way to measure state-making efforts is to include several indicators that would capture the presence of institutions in contested states; several other indicators that would capture the contested states’ ability to assert the monopoly on the use of force; and several other indicators that would capture their administrative and policy implementing capacity. Given that data on contested state are limited and do not allow to capture the ideal measurement of state-making efforts fully, I utilize the available data to attempt to capture three core components of state-making efforts: (i) the institution-building; (ii) the centrality; and (iii) territoriality.

To capture the institutional dimension of the state, I employ the (dfsbuild) indicator, which is an ordinal indicator with four values, between 1 (low state-building efforts) to 4 (very high state-building efforts). To capture the dimension of centrality, I utilize (factions_4_cat) which, just like (dfsbuild), is an ordinal indicator with four values, between 1 (poor levels of centrality, meaning very high levels of factionalism) to 4 (high level of centrality, meaning very low levels of factionalism). Factionalism is a legitimate indicator to capture the ability of contested states to assert monopoly of their rule and act autonomously from the “civil society,” as it speaks of the level of rivalries that may exist in contested states challenging its claims. Thus, both dimensions, the one capturing the level of institutionalization and that of centrality, are given equal weights. To capture the territoriality dimension, I employ (areacontrolcs), which is a continuous variable ranging between 0 and 1. Values closer to 0 represent low territorial outreach of the contested state, while values closer to 1 represent high territorial outreach. After identifying these indicators, one last question that remains to be addressed is what is the best theoretical relationship between the above indicators that would best capture state-making efforts. It should be noted that no agreement exists on whether these components of measuring state strength and/or weakness are to be viewed as analytically distinct. Mann (1984), Hanson and Sigman (2013), and Migdal (1988) see the dimensions of the state as analytically distinct, meaning that they could constitute separate components of the state, which can also be measured separately. Others, like Hanson and Sigman (2013) hold the view that various dimensions of the state are analytically distinct, but that each dimension necessarily affects another, giving room to include these dimensions in a single higher-order conception of state-making efforts. For example, Anderson (1987) conceives of the state as a cumulative variable of each of its dimensions, from “bureaucratic administration, military force, financial resources, territorial integrity, ideological legitimacy, and perhaps others” (1987, 14). In one recent analysis, Fortin-Rittberger (2014) found, for example, that despotic and infrastructural capacities of the state sometimes may be both, positively and negatively related - meaning that sometimes an increase in one may increase the other, while at other times an increase in one may decrease the other. From the perspective of my indicators, this would mean that there may be contested states that may have high institution building, but lack centrality or territorial outreach. In some other cases, contested states may enjoy high levels of each of these dimensions, while others may enjoy low levels of each of these dimensions. Be as it may, I derive the (statemaking) by the following expression:

\[ statemaking = (dfsbuild + factions_{-}4_{-}cat) * areacontrolcs' \]

(’) the prime symbol in the (areacontrolcs) indicator means that any value above 1.00 (100 percent) of territorial control has been recoded to 1.00. This is done because there are only a few cases the are overly stretch above 1.00, providing no useful empirical material for other cases for comparative purposes.

This aggregation captures the state making efforts in a way that it includes the level of institution-building that the contested states have developed, on the one hand, and the level of centrality they have managed to assert, on the other. Additionally, I capture the contested states’ ability to extend these two dimensions over spans of territory, by multiplying the addition of institutionality and centrality with the percent of the territory these entities control.

Note: while useful in capturing the institutional dimension of state-making efforts, (dfsbuild) falls short in capturing other core elements of state making such as contested states’ centrality and territorial reach. Not including some indicators that can capture the centrality and territoriality as the literature on the state suggests, may lead to some anomalous ranking of contested states in their state making efforts. For example, Palestine’s average (dfsbuild) score of 3 out of maximum 4, places it much closer to well-developed entities such as Taiwan and Northern Cyprus, than the in-depth literature on Palestine’s stateness would suggest (Tschirgi 2004; Geldenhuys 2009; Ghanem 2017). Furthermore, the (dfsbuild) alone places contested states such as Somaliland and Transnistria, on average, far below Palestine and much lower than the bulk of literature suggests (Geldenhuys 2009; Blakkisrud and Kolstø 2011; Caspersen 2012; Pegg and Kolstø 2015)

16.6 State-making score - alternative 1 (statemaking_alt_1)

Possible doubts may exist with the theoretical relationship of the three indicators included in the expression. One may legitimately question the use of multiplication of the institutional (dfsbuild) and centrality (factions_4_cat) dimensions to capture the latter’s territorial reach, and thus of state-making efforts overall. I offer an alternative conceptualization where, instead of multiplication, I add the dimension of territoriality with an arithmetic addition. Thus, I give equal weight to all three dimensions of state-making efforts, by adding the 1 to 4 scaled (dfsbuild) and the 1 to 4 scaled (factions_4_cat) and the 1 to 4 scaled (areacontrolcs_4_cat) to capture the alternative measurement model of state-making efforts. The (statemaking_alt_1) is derived from the following expression:

\[ statemaking_{-}alt_{-}1 = dfsbuild + factions_{-}4_{-}cat + areacontrolcs_{-}4_{-}cat \]

16.7 State-making score - alternative 2 (statemaking_alt_2)

Another (simpler) alternative measurement of state making effort I offer is to multiply the (dfsbuild) with the percent of the territorial reach that contested states have. This is in case one would question the inclusion and subsequently the categorization of factions as representing the ability of contested state to to lay claim over the monopoly of power from the center. In this case, the alternative measurement (statemaking_alt_2) is derived from the following expression:

\[ statemaking_{-}alt_{-}2 = (dfsbuild) * areacontrolcs' \] (’) the prime symbol in the (areacontrolcs) indicator means that any value above 1.00 (100 percent) of territorial control has been recoded to 1.00. This is done because there are only a few cases the are overly stretch above 1.00, providing no useful empirical material for other cases for comparative purposes.

16.8 State-making score - alternative 3 (statemaking_alt_3)

One may still question the arithmetic multiplication as a means to theoretically relate the (dfsbuild) and (areacontrolcs), in which case I add both, but use the categorical indicator of control over the territory, to give equal weight to the institutional part of state-making efforts (dfsbuild) and to control over the territory (areacontrolcs_4_cat), both between 1 (low) to 4 (high). Thus, the (statemaking_alt_3) is derived from the following expression:

\[ statemaking_{-}alt_{-}3 = dfsbuild + areacontrolcs_{-}4_{-}cat \]

16.9 Parent instability score (parentinstabl)

Parent state instability can be measured in several ways. One way to measure it is to look at whether or not the parent state has been in conflict with internal or external actors other than the contested state during the existince of the latter, which were born out of the parent state’s territory. When a state faces internal conflicts, such as civil wars, it makes such states unable to effectively direct resources into curbing all the conflicts at the same time, including the one with the contested state. According to Salehyan (2007), when a state undergoes a civil war or other concurrent conflicts, its government is unable to have full control over the territory, and it is a clear indication of institutional weakness. Likewise, the State Failure Task Force (SFTF) suggests that state failure consists of the existence of revolutionary wars, genocides and politicides, and adverse or disruptive regime transitions (for further analysis, see: King and Zeng 2001). Another remote, and rather rudimentary, approach which has been applied in the study of conflicts, is to measure state weakness through the states’ GDP per capita indices (see: Salehyan 2007). While there is some merit to these approaches, I suggest that to capture parent states’ instability, we need to theoretically relate both, the instability that may come due to internal and external conflicts, and their actual weakness or strength. This is because there may be parent/belligerent states, which may face internal or external conflicts with other actors, but, which, at the same time, may be strong in terms of their economic development, the ability to mobilize troops, or simply in terms of military spending - making them able to handle internal conflicts without necessarily facing instability. Think of Israel, which, despite conflicts with various parties in the Middle East, can be considered neither unstable nor weak. At the same time, there can be cases that do not face either internal nor external conflicts but are still too weak to engage the contested state born out of them.

To account for such particularities, I construct a measurement which takes into account both (i) possible internal or external conflicts, and (ii) actual strength and weakness. To capture the first component, I use a continuous indicator (confpdays_r_croot), which registers parent states’ intensity of conflicts with actors other than the contested state during each observed year. The (confpdays_r_croot) includes continuous values ranging between 0 (not involved in any conflict) to 5 (fully involved in conflict). To capture the second component, I use a continuous indicator (troopspcap_r_croot), which registers the parent states’ ability to mobilize troops. The (troopspcap_r_croot) includes values ranging between 0 (very high mobilization capacity) to 5 (very low mobilization capacity). To derive the explanatory variable of parent instability (parentinstabl), I arithmetically add (confpdays_r_croot) and (troopspcap_r_croot). This means that the closer to 0 the (parentinstabl) is, the more stable the parent state is. Further away from zero, the more unstable the parent state is. Thus, the (parentinstabl) is derived from the following expression:

\[ parentinstabl = confpdays_{-}r_{-}croot + troopspcap_{-}r_{-}croot \]

16.10 Parent instability score - alternative 1 (parentinstabl_alt_1)

An alternative to the original measurement model of the explanatory condition of parent state instability is to add the intensity of conflict in which the parent state has been involved (confpdays_r_croot) with the parent state’s military expenditure per capita, instead of with its ability to mobilize troops. Thus, my alternative measurement is derived by adding two equally scaled continuous indicators (confpdays_r_croot) and (milexcap_r_croot). The (parentinstabl_alt_1) is derived from the following expression:

\[ parentinstabl_{-}alt_{-}1 = confpdays_{-}r_{-}croot + milexcap_{-}r_{-}croot \]

16.11 Parent instability score - alternative 2 (parentinstabl_alt_2)

Another alternative that I provide is to add the intensity of conflict in which the parent state has been involved (confpdays_r_croot) with the parent state’s GDP per capita instead. Thus, my second alternative measurement model for the parent state instability is to add (confpdays_r_croot) with (cgdppc_r_croot), an indicator which includes values ranging between 0 (very high GDP per capita) to 5 (very low GDP per capita). The (parentinstabl_alt_2) is derived from the following expression:

\[ parentinstabl_{-}alt_{-}2 = confpdays_{-}r_{-}croot + cgdppc_{-}r_{-}croot \]

16.12 Parent instability score - alternative 3 (parentinstabl_alt_3)

The last alternative I offer is to simply keep the measurement of parent state instability with its number of days in conflict with internal and/or external actors, other than with the contested state. The (parentinstabl_alt_3) is derived from the following expression:

\[ parentinstabl_{-}alt_{-}3 = confpdayswoCS \]

16.13 International legitimacy score (intlegit)

The level of contested states’ international legitimacy is measured by the ratio of the number of legal recognitions these entities have accumulated against the total number of UN member states in any given year. Martin Wight defines international legitimacy as “the collective judgment of international society about rightful membership of the family of nations; how sovereignty may be transferred; how state succession should be regulated, when large states break up into smaller, or several states combine into one” (1972, 1). The aspect of collective judgment about entities’ rightful membership in the family of states has also been specified elsewhere. Shane Mulligan suggests that international legitimacy is not a condition that might be perceived, but “a matter of conscious collective decision” (2006, 362). Contested state’s by their nature of existence as de facto separate entities lack the full collective recognition by the existing society of states. However, their legitimacy, as partial as it may be, can be measured by looking at the number of the existing members of the society of states (UN member states) that have individually granted these entities recognition as rightful members of the society of states. Thus, the (intlegit) is derived by the following expression:

\[ intlegit = = \frac {recog}{unmemtot}\]

16.14 Time in peace score (tpeace)

The explanatory condition of time in peace is measured by the ratio of cumulative days the contested state has spent in peace and their total days of existence. For example, if a contested state has spent 90 days in conflict with its parent during the entire calendar year, it means that it has spent 25 percent of its time at war with the parent during a given year. Thus its time in peace has been 75 percent of the time or 0.75 ratio of cumulative days in peace against 0.25 at war. The (tpeace) is derived by the following expression:

\[ tpeace = 1-(confcsdcum / dayscum)\]

16.15 Time in peace score (tpeace_alt_1)

The alternative to (tpeace) is a very similar indicator; however, when deriving the mean value of the time-series data, the mean of cumulative is replaced by the actual cumulative time in peace at the end of case observation. The (tpeace_alt_1) is derived by the following expression:

\[ tpeace_{-}alt_{-}1 = 1-(confcsdcum / dayscum)'\]

(’) the prime symbol means that when deriving the mean value of the time-series data for each case, the actual mean is not the mean value of the cumulative, but the actual cumulative time in peace at the end of case observation.

References

Anderson, Lisa. 1987. “The State in the Middle East and North Africa.” Comparative Politics 20 (1): 1–18. https://doi.org/10.2307/421917.

Blakkisrud, Helge, and Pål Kolstø. 2011. “From Secessionist Conflict Toward a Functioning State: Processes of State- and Nation-Building in Transnistria.” Post-Soviet Affairs 27 (2): 178–210. https://doi.org/10.2747/1060-586X.27.2.178.

Bolt, Jutta, Robert Inklaar, Herman de Jong, and Jan Luiten van Zanden. 2018. “Rebasing ‘Maddison’: New Income Comparisons and the Shape of Long-Run Economic Development.” Research Memorandum 174. Maddison Project Database, Version 2018. Groningen: University of Groningen: Groningen Growth and Development Centre. https://www.rug.nl/ggdc/html_publications/memorandum/gd174.pdf.

Bureau of Labor Statistics. 2019. “Databases, Tables & Calculators by Subject: CPI-All Urban Consumers (Current Series).” United States Department of Labor. 2019. https://data.bls.gov/pdq/SurveyOutputServlet.

Byman, Daniel, Peter Chalk, Bruce Hoffman, William Rosenau, and David Brannan. 2001. “Trends in Outside Support for Insurgent Movements.” Product Page. Santa Monica, CA: RAND National Security Research Division. https://www.rand.org/pubs/monograph_reports/MR1405.html.

Carter, David B. 2012. “A Blessing or a Curse? State Support for Terrorist Groups.” International Organization 66 (1): 129–51. https://doi.org/10.1017/S0020818311000312.

Caspersen, Nina. 2012. Unrecognized States: The Struggle for Sovereignty in the Modern International System. Cambridge, UK; Malden, M.A.: Polity Press.

CIA. 2019. “The World Factbook.” 2019. https://www.cia.gov/library/publications/the-world-factbook/.

Coggins, Bridget. 2011. “Friends in High Places: International Politics and the Emergence of States from Secessionism.” International Organization 65 (3): 433–67. https://doi.org/10.1017/S0020818311000105.

Cunningham, Kathleen Gallagher. 2013. “Actor Fragmentation and Civil War Bargaining: How Internal Divisions Generate Civil Conflict.” American Journal of Political Science 57 (3): 659–72. http://www.jstor.org/stable/23496645.

Desch, Michael C. 1996. “War and Strong States, Peace and Weak States?” International Organization 50 (2): 237–68. http://www.jstor.org/stable/2704078.

Esfahani, Hadi Salehi, and M. Hashem Pesaran. 2009. “The Iranian Economy in the Twentieth Century: A Global Perspective.” Iranian Studies 42 (2): 177–211. http://www.jstor.org/stable/25597543.

Florea, Adrian. 2014. “De Facto States in International Politics (1945 - 2011): A New Data Set.” International Interactions 40 (5): 788–811. https://doi.org/10.1080/03050629.2014.915543.

———. 2017. “De Facto States: Survival and Disappearance (1945–2011).” International Studies Quarterly, March. https://doi.org/10.1093/isq/sqw049.

Fortin-Rittberger, Jessica. 2014. “Exploring the Relationship Between Infrastructural and Coercive State Capacity.” Democratization 21 (7): 1244–64. https://doi.org/10.1080/13510347.2014.960207.

Geldenhuys, Deon. 2009. Contested States in World Politics. Basingstoke, UK: Palgrave.

Ghanem, As’ad. 2017. “Historical Background of the State-Building Process in Palestine.” In Between State and Non-State: Politics and Society in Kurdistan-Iraq and Palestine, edited by Gülistan Gürbey, Sabine Hofmann, and Ferhad Seyder, 43–58. New York: Palgrave MacMillan.

Gleditsch, Kristian S., and Michael D. Ward. 1999. “A Revised List of Independent States Since the Congress of Vienna.” International Interactions 25 (4): 393–413. https://doi.org/10.1080/03050629908434958.

———. 2016. “Measuring Space: A Minimum-Distance Database and Applications to International Studies:” Journal of Peace Research, July. https://doi.org/10.1177/0022343301038006006.

Gleditsch, Nils Petter, Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg, and Håvard Strand. 2002. “Armed Conflict 1946-2001: A New Dataset.” Journal of Peace Research 39 (5): 615–37. https://doi.org/10.1177/0022343302039005007.

Grävingholt, Jörn, Sebastian Ziaja, and Merle Kreibaum. 2015. “Disaggregating State Fragility: A Method to Establish a Multidimensional Empirical Typology.” Third World Quarterly 36 (7): 1281–98. https://doi.org/10.1080/01436597.2015.1038340.

Hafner-Burton, Emilie M., and Alexander H. Montgomery. 2006. “Power Positions: International Organizations, Social Networks, and Conflict.” Journal of Conflict Resolution 50 (1): 3–27. https://doi.org/10.1177/0022002705281669.

Hanson, Jonathan K., and Rachel Sigman. 2013. “Leviathan’s Latent Dimensions: Measuring State Capacity for Comparative Political Research.” SSRN Scholarly Paper ID 1899933. Rochester, NY: Social Science Research Network. http://papers.ssrn.com/abstract=1899933.

Harbom, Lotta, Erik Melander, and Peter Wallensteen. 2008. “Dyadic Dimensions of Armed Conflict, 1946—2007.” Journal of Peace Research 45 (5): 697–710. https://doi.org/10.1177/0022343308094331.

Högbladh, Stina, Therése Pettersson, and Lotta Themnér. 2011. “UCDP External Support – Primary Warring Party Dataset.” In. Montreal, Canada: Swedish National Data Service. https://snd.gu.se/en/catalogue/study/ext0034.

ICJ. 2018. “Press Releases.” International Court of Justice. 2018. https://www.icj-cij.org/en/press-releases.

IGAD. 2019. “The IGAD Region.” 2019. https://igad.int/index.php/about-us/the-igad-region.

King, Gary, and Langche Zeng. 2001. “Improving Forecasts of State Failure.” World Politics 53 (04): 623–58. https://doi.org/10.1353/wp.2001.0018.

Kreutz, Joakim. 2010. “How and When Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset.” Journal of Peace Research 47 (2): 243–50. https://doi.org/10.1177/0022343309353108.

Mann, Michael. 1984. “The Autonomous Power of the State: Its Origins, Mechanisms and Results.” European Journal of Sociology / Archives Européennes de Sociologie 25 (02): 185–213. https://doi.org/10.1017/S0003975600004239.

Migdal, Joel S. 1988. Strong Societies and Weak States: State-Society Relations and State Capabilities in the Third World. Princeton, N.J.: Princeton University Press.

Mulligan, Shane P. 2006. “The Uses of Legitimacy in International Relations.” Millennium 34 (2): 349–75. https://doi.org/10.1177/03058298060340021801.

Pegg, Scott, and Pål Kolstø. 2015. “Somaliland: Dynamics of Internal Legitimacy and (Lack of) External Sovereignty.” Geoforum 66: 193–202. https://doi.org/http://dx.doi.org/10.1016/j.geoforum.2014.09.001.

Pettersson, Therése, Stina Högbladh, and Magnus Öberg. 2019. “Organized Violence, 1989–2018 and Peace Agreements.” Journal of Peace Research 56 (4): 589–603. https://doi.org/10.1177/0022343319856046.

Saideman, Stephen M. 2001. The Ties Divide: Ethnic Politics, Foreign Policy, and International Conflict. New York: Columbia Univ. Pr.

Salehyan, Idean. 2007. “Transnational Rebels: Neighboring States as Sanctuary for Rebel Groups.” World Politics 59 (2): 217–42. http://www.jstor.org/stable/40060187.

———. 2008. “No Shelter Here: Rebel Sanctuaries and International Conflict.” The Journal of Politics 70 (1): 54–66. https://doi.org/10.1017/S0022381607080048.

Salehyan, Idean, Kristian Skrede Gleditsch, and David E. Cunningham. 2011. “Explaining External Support for Insurgent Groups.” International Organization 65 (4): 709–44. http://www.jstor.org/stable/23016231.

Sawyer, Katherine, Kathleen Gallagher Cunningham, and William Reed. 2017. “The Role of External Support in Civil War Termination.” Journal of Conflict Resolution 61 (6): 1174–1202. https://doi.org/10.1177/0022002715600761.

Schultz, Kenneth A. 2010. “The Enforcement Problem in Coercive Bargaining: Interstate Conflict over Rebel Support in Civil Wars.” International Organization 64 (2): 281–312. http://www.jstor.org/stable/40608016.

Singer, J. David. 1988. “Reconstructing the Correlates of War Dataset on Material Capabilities of States, 1816–1985.” International Interactions 14 (2): 115–32. https://doi.org/10.1080/03050628808434695.

Singer, J. David, Stuart Bremer, and John Stuckey. 1972. “Capability Distribution, Uncertainty, and Major Power War, 1820-1965.” In Peace, War, and Numbers. Edited by Bruce M. Russett., edited by Bruce M Russett, 19–48. Beverly Hills [Calif.: Sage Publications.

SIPRI. 2018. “SIPRI Military Expenditure Database (1949 - 2017).” Stockholm International Peace Research Institute. 2018. https://www.sipri.org/databases/milex.

Skocpol, Theda. 1985. “Bringing the State Back in: Strategies of Analysis in Current Research.” In Bringing the State Back in, edited by Peter B Evans, Dietrich Rueschemeyer, and Theda Skocpol, 3–37. Cambridge, UK: Cambridge University Press.

Tschirgi, Dan. 2004. “Palestine 2003: The Perils of de Facto Statehood.” In De Facto States: The Quest for Sovereignty, edited by Tozun Bahcheli, Barry Bartmann, and Henry Srebrnik, 187–209. Oxon, UK; New York, N.Y.: Routledge.

United Nations. 2018a. “Official Document System of the United Nations.” Official Document System. 2018. https://documents.un.org/prod/ods.nsf/home.xsp.

———. 2018b. “Resolutions.” United Nations Security Council. 2018. https://www.un.org/securitycouncil/content/resolutions-0.

———. 2018c. “Resolutions.” United Nations General Assembly. 2018. https://www.un.org/en/sections/documents/general-assembly-resolutions/.

———. 2018d. “United Nations Member States.” 2018. https://www.un.org/en/member-states/.

———. 2018e. “United Nations Peace Operations.” 2018. https://www.unmissions.org/.

Weber, Max. 2004. “Politics as a Vocation.” In The Vocation Lectures, edited by David S Owen and Tracy B Strong, 32–94. Indianapolis, Ind.; Cambridge, N.Y.: Hackett Pub.

Wickham, Hadley. 2019. Wrappers Around the ’Xml2’ and ’Httr’ Packages to Make It Easy to Download, Then Manipulate, HTML and XML. http://rvest.tidyverse.org/.

Wight, Martin. 1972. “International Legitimacy.” International Relations 4 (1): 1–28. https://doi.org/10.1177/004711787200400101.

World Bank. 2019a. “GDP Per Capita, PPP (Constant 2011 International $).” The World Bank Group. 2019. https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.KD?locations=PG.

———. 2019b. “GDP Per Capita (Current US$).” The World Bank Group. 2019. https://data.worldbank.org/indicator/NY.GDP.PCAP.CD.

———. 2019c. “GDP Per Capita Growth (Annual %).” The World Bank Group. 2019. https://data.worldbank.org/indicator/NY.GDP.PCAP.KD.ZG.

———. 2019d. “Military Expenditure (Current USD).” The World Bank Group. 2019. https://data.worldbank.org/indicator/MS.MIL.XPND.CD.

Codebook for the Dataset on Contested States in World Politics, 1945 - 2017

Sh. K.

12 April 2021

1 The reader’s guide to the dataset (and codebook)

1.1 Structure of the codebook

1.2 Indicator lots

1.3 Rules for naming indicators

1.4 Data sources and formats

1.5 Cautionary notes

2 PART I: Case identifiers

2.1 Unique identifier for each case-year observation of all contested states (id1) | (A)

2.2 Unique identifier for each case-year observation of each contested state (id2) | (A)

2.3 Unique identifier for each contested state (id3) | (A)

2.4 Contested state names (namecs) | (A)

2.5 Contested state code (dfscode) | (C)

2.6 Contested state abbreviation (dfsabbrev)| (C)

2.7 Parent state name (namep) | (C)

2.8 Parent state code (ccode) | (C)

2.9 Parent state abbreviation (abbrp) | (C)

3 PART II: Lifecycle indicators

3.1 Contested state declaration of independence (indpndcecs) | (A)

3.2 Contested state start date (csstart) | (A)

3.3 Contested state end date (csend) | (A)

3.4 Contested state observation years (yearcs) | (A)

3.5 Contested states in days (days) | (D)

3.6 Contested states cumulative days of existence (dayscum) | (D)

3.7 Contested states in months (months) | (D)

3.8 Contested states cumulative months of existence (monthscum)| (D)

3.9 Contested states in years (years) | (D)

3.10 Contested states cumulative years of existence (yearscum) | (D)

3.11 Parent state UN membership (unp) | (A)

3.12 Parent states cumulative days of existence (dayscump) | (D)

3.13 Parent states cumulative months of existence (monthscump) | (D)

3.14 Parent states cumulative years of existence (yearscump) | (D)

3.15 Age difference in days (diffd) | (D)

3.16 Age difference in months (diffm) | (D)

3.17 Age difference in years (diffy) | (D)

3.18 Contested state outcome (outcome) | (A)

3.19 Contested state sub-outcome (suboutcome) | (A)

3.20 Maturity of the parent state (newstate)| (A)

4 PART III: Stateness indicators

4.1 Contested state passports (passportcs) | (A)

4.2 Parent state passports (passportp) | (A)

4.3 Third-party passports (passportes) | (A)

4.4 Contested state central bank (cbank) | (A)

4.5 Contested state currency (currencycs) | (A)

4.6 Parent state currency (currencyp) | (A)

4.7 Third-party currency (currencyes) | (A)

4.8 Control over the capital city (capitalcontrol) | (A)

4.9 Contested state ministry of foreign affairs (mfa) | (A)

4.10 State-building efforts by the contested state (dfsbuild) | (C)

4.11 Number of factions in contested states (factions) | (C)

5 PART IV: Territorial indicators

5.1 Total official area of contested states (areacs) | (A)

5.2 Total official area of parent states (areap) | (A)

5.3 Approximate controlled area by contested states (controlcs) | (A)

5.4 Approximate area controlled by parent states (controlp) | (A)

5.5 Ratio of area under the control of contested states (areacontrolcs) | (D)

5.6 Ratio of area under the control of parent states (areacontrolp) | (D)

5.7 Ratio of the official size of the area of contested states against their parent states (arearatio) | (D)

5.8 Ratio of the controlled size of the area by contested states against their parent states (controlratio) | (D)

5.9 Contested states’ access to sea (seacs) | (A)

5.10 Parent states’ access to sea (seap) | (A)

5.11 Contested states as islands (islandcs) | (A)

5.12 Parent states as islands (islandp) | (A)

5.13 Contested states’ territorial contiguity (contigcs) | (A)

5.14 Parent states’ territorial contiguity (contigp) | (A)

5.15 Contested states’ land border with their parent states (borderwp) | (A)

5.16 Contested states’ approximate land border length with their parent states (blengthwp) | (A)

5.17 Contested states’ land borders with entities other than their parent states (bordero) | (A)

5.18 Contested states’ land border with external supporters (borderwes) | (A)

5.19 Distance (km): capital cities of contested states and their parent states (distcapkm) | (A)

5.20 Distance (time): capital cities of contested states and their parent states (distcaphrs) | (A)

5.21 Distance (km): capital cities of contested states and the border with their parent states (distbcapkm) | (A)

5.22 Distance (time): capital cities of contested states and the border with their parent states (distbcaphrs) | (A)

5.23 Contested states’ terrain advantage (terrain) | (A)

6 PART V: Demographic indicators

6.1 Population of contested states (popcs) | (A)

6.2 Population of parent states (popp) | (A)

6.3 Real population of parent states (poppwocs) | (A)