Introduction

The death penalty is the ultimate form of punishment as the government ends the life of a person. The punishment is also unique in that it is irreversible; wrongly accused and convicted persons can never be compensated. It is therefore not surprising that the death penalty has become the most politically controversial form of punishment. For centuries, political actors and parties, various interest groups, and international organizations have been engaged in long-standing conflicts over the use of the death penalty. Overall, these conflicts have resulted in significant variations of death penalty usage between and within countries but also over time. The general global trend since 1863, when Venezuela became the first country to abolish the death penalty for all crimes, has been towards the abolition of the death penalty.Footnote 1 Throughout the decades, both the number and share of countries using the death penalty has decreased. However, there are still many countries that apply the death penalty and some countries even have reintroduced it after lengthy periods of abolition.

Although the death penalty is politically controversial, surprisingly, few studies in political science have examined the determinants of death penalty usage in the world. In the few studies that have been undertaken (notably Anckar 2004; Greenberg and West 2008; Kim 2016; MacGann and Sandholtz 2012; Mathias 2013; Neumayer 2008a; Neumayer 2008b; Suh 2015), authors have generally focused on cross-national differences during the time period 1950s/1960s and onwards.

During the last few decades, a number of social science databases have been released. Although some of the databases, e.g. the Varieties of Democracy (V-dem) dataset, contain yearly observations for a wide variety of data over extensive periods, data on the worldwide use of capital punishment have not been included among the variables. This is probably a reflection of the fact that the global usage of capital punishment has interested criminologists and scholars of international law to a much higher extent than social scientists (e.g. Schabas 1997; Hood and Hoyle 2015). Yet, the death penalty has many social and political implications. For instance, until quite recently, serious contenders for the US presidency could not afford to take a stand against the death penalty (e.g. Wilkins 2007, p 803); the racial and gender bias associated with the death penalty in the USA is also widely acknowledged and discussed (e.g. Harmon et al. 2023; Robertson and Bell 2022; Ramirez 2021); and the abolition of the death penalty in Eastern Europe is largely explained by the fact that Western European countries made abolition a requirement to join the Council of Europe and the European Union (Fawn 2001). Here, the ambition is to collate information on death penalty usage into one dataset, comprising all currently independent countries of the world from the year 1800 and onwards. Although information on death penalty usage is provided in a number of sources on the internet, the sources do not contain adequate information on how death penalty policies have changed over time in the countries of the world.Footnote 2 For instance, in many countries, notably in Latin America, the death penalty has been abolished and reinstated at various points in time. Moreover, existing sources do not contain information about the extent to which countries have experienced long periods of de facto moratoriums on executions.

The present article introduces the Comparative Death Penalty Database (CDPD). The aim of the database is to improve the conditions for comparative studies of the development of capital punishment over time. The database provides data on 12 variables related to death penalty usage during the time period 1800–2022. Additionally, the database makes use of a modified categorization that differentiates death penalty usage in a new way. It includes more than 19,000 country-year observations for all currently independent states. This material allows researchers to study not only the development of death penalty usage over time but also the diffusion of death penalty policies between countries.

In the next section, we present a categorization of death penalty usage. We then provide information about the structure of the database (variables) and the material used when compiling the database. We then proceed by presenting some descriptive analyses of death penalty trends over time and by illustrating the relevance of the database for comparative research. The last section offers some conclusions and suggestions for future research.

Concept and categorization

The database is based on a general and conventional definition of the death penalty. Death penalty (or capital punishment) refers to the execution of a state-sanctioned practice of killing a person as a punishment for a crime that a juridical (rule-governed) process has concluded that the person in question has conducted. According to this definition, the death penalty is exclusively related to the justice system within countries, thereby excluding, for instance, government sanctioned arbitrary killings and political murders. Additionally, it distinguishes between executions and death sentences. A death sentence constitutes a necessary but not sufficient condition for an execution, as many death sentences are never carried out.

When measuring death penalty usage, several options are available. One alternative is to operate with a dichotomous variable, where countries that actually execute persons are confined to one category and all other countries to another. Such a categorization would create a category that contains considerable variations. For example, countries that have abolished the death penalty for all crimes would fall into the same category as countries that have death penalty statutes but where death sentences are not carried out. A second alternative is to use a quantitative measure of death penalty usage that pays regard to the number of crimes punishable by death or the number of executions undertaken by countries each year. However, the number of crimes that are punishable by death does not indicate the actual use of the death penalty and the data on executions tend to be highly unreliable. Accordingly, very sophisticated measures of death penalty usage cannot be applied in a global context. The problem is aggravated when the ambition of the database is to cover a long time period and all countries in the world.

Most studies in the field therefore use a third alternative, offered by Amnesty International. This classification scheme consists of four categories (e.g. Hood and Hoyle 2015):

  • Abolitionist for all crimes: countries that have abolished the death penalty for all crimes under all circumstances.

  • Abolitionist for ordinary crimes only: countries use the death penalty only for exceptional crimes (e.g. crimes under military law or committed in exceptional circumstances such as wartime).

  • Abolitionist de facto: countries have the death penalty but have not executed anyone during the past ten years or have made international commitments not to execute death sentences.

  • Retentionist: countries that allow the death penalty and have been carrying out executions within the past ten years.

In the present contribution, we apply Amnesty International’s classification, albeit with three modifications. First, countries with death penalty statuses are considered de facto abolitionist only if they have not executed anybody for a period of at least ten years. In other words, international commitments not to execute death sentences are disregarded. This decision is based primarily on the fact that international commitments to abolish the death penalties did not exist prior to the adoption of the International Covenant on Civil and Political Right from 1966. Moreover, the covenant did not demand full abolition of the death penalty, but instead required the parties to restrict the use of capital punishment ‘to the most serious crimes in accordance with the law’ (art. 6.2). Not until 1989, when the United Nations, by resolution 44/128, adopted the Second Optional Protocol to the International Covenant on Civil and Political Rights, aiming at the abolition of the death penalty, was there an international treaty that required states to abolish the death penalty altogether (Neumayer 2008a, pp. 3–4).

Second, we split the second category (Abolitionist for ordinary crimes only) into two categories. In some of these countries, death sentences have been carried out during the last ten years, and these cases are referred to a separate category, labelled Abolished for ordinary crimes only but used during the last ten years. The main reason for this decision is to make the situation comparable to the distinction made by retentionist countries (where executions actually occur) and de facto abolitionist countries (where a de facto moratorium exists). It is reasonable to consider a country that has executed a person for crimes committed under exceptional circumstances as more positive towards the death penalty than countries that have not committed executions under such circumstances. Needless to say, researchers who need or want to use the original categorization can combine the two categories.

Third, for newly born states with death penalty statutes, the question arises if they should be classified as de facto abolitionist or retentionist at the time of their independence. Simply accounting for whether executions have occurred on the territory during the last decade under colonial rule is not necessarily the best solution, as the decision to execute persons ultimately was taken by a foreign power, perhaps against the traditions and/or the will of the local population. On the other hand, it is equally possible that the colonial power had a more restricted view of the death penalty than the local population in general, and instead acted as a restraining force against more active use of the death penalty by the local population. Therefore, new states that had death penalty statutes at the time of independence are considered de facto abolitionists at that point in time only insofar as it has been possible to verify that more than ten years had elapsed since the last execution took place on their territories and if no executions occurred during the first ten years of their independence.

The categorization is treated as an ordinal scale of death penalty usage. Abolitionist countries constitute one endpoint, followed by countries abolitionist for ordinary crimes only, countries abolitionist for ordinary crimes only but where at least one execution has occurred in the last 10 years, countries that are abolitionist de facto, and, finally, retentionist countries.

Data sources

Collecting data on death penalty regimes is a challenging undertaking and we have used various sources of information to control and quality-assure information. The United Nations has conducted regular surveys on death penalty usage worldwide since 1975, and these reports have been beneficial for compiling data for the last half-century (United Nations, Capital Punishment, Report of the Secretary General, 1975 E/5616–A/HRC/45/20/: Report of the Secretary General on the Question of the Death Penalty). Since 1977, Amnesty International has campaigned against the use of capital punishment and published statistics on how the death penalty is applied in the world (Amnesty International, Annual Reports 1976–2023). Although these reports are highly reliable for determining whether the death penalty is abolished or not de jure, they are less reliable when it comes to determining the number of executions that have occurred, especially during the 1970s and 1980s. In addition to the sources above, we have used data provided by Roger Hood, Caroline Hoyle, and William Schabas in their seminal studies on the application of the death penalty in the world (Hood and Hoyle 2015; Schabas 1997) but also consulted Deets (1947–1948) and Patrick (1965). Finally, Wikipedia’s internet site ‘Capital punishment by country’ has been extremely useful.Footnote 3

However, in general, the farther back in time we go, the more difficult it is to get reliable information, particularly on whether executions have occurred during the latest ten years or not, but also, surprisingly, on the formal status of the death penalty. This challenge is especially the case for Latin America in the late 19th and early twentieth centuries when some countries alternately abolished and reinstated the death penalty. Here, several works on the death penalty in the respective countries have been extremely useful, notably (Baeza Muños 2016; Castro 2021; Engstrom 2013; Hurtado Pozo 1979; Moreno Martinez 2011; Ochoa-Luis 2017). In addition, we have relied on different versions of the constitutions of the countries. Indeed, some Latin American countries changed their constitutions quite often in this period. For instance, Venezuela had no less than 20 constitutions from 1811 to 1936, whereas Peru had ten constitutions between the years 1823 to 1933, of which the constitutions of 1856, 1860, and 1867 changed the status of the death penalty. A comprehensive list of the sources used is provided in the codebook of this dataset.

The most challenging task has been separating de facto abolitionist and retentionist countries. Many Latin American countries were also restrictive in carrying out executions under certain periods when death penalty statutes were in force. As data on executions is scarce and unreliable, it is sometimes quite difficult to reach definitive conclusions about the extent to which countries in the region have experienced de facto moratoriums of ten years or more in the late 1800s or early 1900s. Additionally, in some cases, it has been a challenge to collect information about historical states that no longer exist. For example, there are missing data for some of the independent states that were integrated into the unified Germany or Italy in the 1800s.

Structure of database

The database includes data for 206 independent countries. Except for some historical states that no longer exist, all independent states are part of the database since 1800 or the year they received their independence. Since the year 1800, the composition of the international state system has changed as some states have dissolved, and new states have become independent. This development has also affected global and regional compositions of death penalty usage. Hence, changes in death penalty status on aggregated levels can reflect changes in the international state system rather than changes of death penalty usage among older states. Furthermore, as the database reflects country-level observations, subnational differences that, in some cases, are quite significant are not reflected in the data. In some federal countries, the choice to either apply the death penalty or not is left to the states. This means that the death penalty is allowed in some states but forbidden in others. This is obviously the case in the USA, but historically, there are many examples of federations where death penalty usage has varied between the federative units. Since, in the present dataset, the unit of analysis is the national level, we have classified federal countries as abolitionist only insofar as the death penalty is forbidden throughout the federation.

The main variable of the database is the status of death penalty usage. This ordinal variable reflects the use of the death penalty for each country and year on December 31. The variable has five categories: Abolitionist for all crimes, Abolitionist for ordinary crimes only, Abolished for ordinary crimes only but used during the previous ten years, Abolitionist de facto, and Retentionist. To allow for indications of change, the database includes a variable that lags the status of death penalty usage with one year. Based on a comparison of the two status variables, a third variable flags whether a change of the status between two years has occurred or not. A fourth variable, finally, accounts for the direction of change, i.e. if the country has moved towards a more positive or negative position to the death penalty along the ordinal scale applied.

The observations are structured according to country codes from the Correlates of War project (https://correlatesofwar.org/data-sets/cow-countrycodes-2/) and to the V-dem database programme's modification of these codes (v-dem.net). As these code systems have become standard among international databases, merging our variables with other databases and adding variables from external datasets is relatively easy.

Descriptive presentation

In Table 1, we present basic descriptions of the variables in the database. The database comprises 19,204 country-year observations in 206 countries from 1800 to 2022. The use of the death penalty has dominated this period. Most of the cases (60 per cent) are placed in the category where  the death penalty is in use, while a minority (20 per cent) of the cases are registered as  completely abolitionist. Altogether, we have registered 370 changes in the death penalty status over the period. In 285 cases, the death penalty policy has become more restricted, whereas it has become less restricted in 85 cases. As shown in Table 2, the popularity of the death penalty varies heavily in time, but there is a clear trend towards abolition over time. Yet, until the 1970s, the death penalty was widely applied in the world. In the mid 1970’s the abolitionist movement got under way and accelerated strongly during the 1980’s. Since then, the share of abolitionist countries has grown at a steady rate at the same time as the share of retentionist countries has decreased markedly. Furthermore, the group of countries that still have death penalty statutes but where executions are not carried out has grown. Since the year 2000, a majority of the cases are abolitionist for ordinary crimes.

Table 1 Descriptive statistics 1800–2022
Full size table
Table 2 Death penalty status 1800–2022 (per cent)
Full size table

Although the long-term trend is towards abolitionism, the status of the death penalty is, of course, normally stable between the two years. As we recall, there are in total 370 cases of changes, making up less than two per cent of all cases. Overall, the most stable categories are found at the two endpoints of the ordinal scale, meaning that countries that are either abolitionist or retentionst are the least likely to change their death penalty status. The least stable category, again, is the one made up of countries that have abolished the death penalty for ordinary crime only but used it during the  previous ten years. It is also quite common that countries move directly from the endpoint retentionist to abolitionist, as this pattern is observed in 43 cases. Furthermore, there are 36 cases of movement from the category abolitionist in practice to abolitionist and 34 cases from the category abolitionist for ordinary crimes only to the category abolitionist.

Based on the findings in Table 1, it is not surprising to find that movements towards abolition are far more frequent than movements towards a reintroduction of the death penalty. Altogether there are 285 cases of movements towards abolition, which accounts for 77 percent of all cases where a change in the death penalty status has occurred. However, only a minority of the cases (113) have resulted in the complete abolition of the death penalty. The other 172 cases reflect movements towards, but not to complete, abolition.

Although movements towards a more positive view of the death penalty are less frequent, they are, by no means, rare. A total of 85 cases are found in the material. More than half of the cases reflect a situation where countries have resumed executions after more than a decade of de facto abolitionism. In yet another ten cases, countries have resumed executions for crimes committed in times of war. Finally, it is worth noting that only thirteen countries have moved directly from the category completely abolished to retentionist.

In sum, the statistical overview strongly reflects the general global development towards a more negative view of the death penalty. However, it is worth emphasizing that this is not a unidirectional trend; the evidence clearly shows that countries are prepared to change their death penalty policies in the other direction as well.

Discussion and conclusion

In this article, we introduce the CDPD dataset, which can contribute to research on the death penalty usage in several ways. First, the dataset allows researchers to study the global development of the death penalty from 1800 and onwards, whereas studies up to now have focused mainly on developments since 1950. As shown in our statistical presentations above, the usage of the death penalty has varied extensively prior to the mid-twentieth century. Previous studies have identified a number of plausible explanations of death penalty usage. These include the level of democracy (Brettschneider 2002; Burt 1994, 90; Neumayer 2008a; Ruddell and Urbina 2004; Greenberg and West 2008; Neumayer 2008b; McGann and Sandholtz 2012; Mathias 2013; Kim (2016), diffusion (Neumayer 2008a; 2008b; McGann and Sandholtz 2012), socioeconomic development (Mathias 2013; Kim 2016), state power (Kim 2016), law tradition (Greenberg and West 2008; Neumayer 2008b), and religion (Mathias 2013; McGann and Sandholtz 2012; Neumayer 2008b; Suh 2015). The present dataset makes it possible to study to what extent the relationships between these variables and death penalty usage varies over time and space. Accordingly, for the first time, researchers interested in the death penalty can explore such questions as how regime change affects a country's use of the death penalty, how the death penalty has been employed in different time frames, or how geographical or political variables have affected countries' use of the death penalty over time.

Second, based on the categorization applied in the database, it is possible to conduct more detailed studies of death penalty trends than hitherto. Although we are, undoubtedly, witnessing a general trend towards abolition, the trend is neither linear nor straightforward; some countries have abolished the death penalty abruptly whereas the process in other countries has been slow and stepwise. It is also noteworthy that a considerable number of states have reintroduced the death penalty at some point in time. The database makes it possible to capture such trends and developments. Finally, the database provides researchers with an opportunity to compare if changes between categories along the ordinal scale are affected by different sets of variables.

We plan to regularly update the database. Additionally, we work to complete missing data for some historical states. Updating historical states in the 1800’s is a challenging and time-consuming work, which requires access to historical documentation. By encouraging researchers to complement and correct figures in the database, we strive to ensure that figures are reliable and that changes in the status of the death penalty are reflected in the data. The dataset will be updated every four years and the new versions of the dataset will be freely available. The death penalty is an under-researched area in political science. We hope that the introduction of the database will constitute an important step in the process of developing a comprehensive theory about the determinants of the death penalty. In addition, it allows researchers to also study consequences of death penalty usage in different geographical and cultural contexts and in different time periods.