- Widely-Used Compendia of Development Data
- Health and Health Care Data
- Data for Latin America and the Caribbean
- Infant, Child, and Maternal Mortality
- Family Planning
- Vaccines and Immunizations
- Economic Affluence
- Educational Attainment
- Income Inequality
- Income Pover
- Water and Sanitation
- Geographical Variables
- Democracy, Constitutions, Civil and Political Rights, Women in Parliament
- State Capacity
- Free-Market Orientation
1. Widely-Used Compendia of Development Data
OurWorldinData.org compiles a very wide range of data from highly reputable sources. Here is their Covid-19 Data Explorer. You will have to look around the vast website to see what they have.
The World Bank World Development Indicators (“WDI”) are the most widely-cited data pertaining to economic and social development. The WDI include hundreds of variables pertaining to GDP per capita, income inequality, income poverty, mortality (adult, infant, maternal), life expectancy at birth, age at first marriage, fertility, population in different age groups, urban population, population density, contraceptive prevalence, birth attendance, doctors per capita, nurses per capita, hospital beds per capita, access to sanitation and safe water, immunization rates, adult illiteracy rates, HIV prevalence, etc.. Some indicators are disaggregated by gender, urban/rural, and so on. To use the WDI statistical compiler, (1) select a country or countries, then click “next”; (2) select variables (series), then click “next”; and (3) select years, then click “apply changes.”
The Quality of Government datasets housed at the University of Gothenburg, Sweden (but published in English), consists of about 2500 indicators on national quality of governance (as measured by such indicators as corruption, bureaucratic quality, democracy), some of its hypothesized causes (e.g., colonial origin, religion, and ethnolinguistic fractionalization), and some of its hypothesized consequences (e.g., GDP per capita, educational attainment, infant mortality, gender bias, environmental sustainability, life satisfaction, and trust). The indicators were compiled from about 100 sources. The Standard TS [time series] dataset has data on each indicator in each year from 1946 to 2016, provided a credible estimate can be found. The unit of analysis is the country-year — hence, Sweden-1946, Sweden-1947, and so on.
UNdata includes 33 regularly updated databases with more than 60 million separate pieces of information on such topics as “Agriculture, Crime, Education, Employment, Energy, Environment, Health, HIV/AIDS, Human Development, Industry, Information and Communication Technology, National Accounts, Population, Refugees, Tourism, and Trade.”
The Human Development Reports published annually by the United Nations Development Programme (UNDP) include a wide range of statistical data related to people’s ability to live a long and healthy life, acquire knowledge, and achieve a decent standard of living. Besides the Human Development Report Office’s annual global Reports covering all countries, associated agencies in most countries of the world have published national and subnational Human Development Reports. The Human Development Report 2010, the 20th Anniversary Edition, reaffirms the benefits of the whole 20-year project and tweaks significantly, with careful justification, the algorithms for calculating the major indices, including the Human Development Index. By 2015, however, some the 2010 revisions had been re-thought. A comparison of the technical notes of the 2013 and 2015 reports shows that in 2015 the two education sub-indices were combined using the arithmetic rather than the geometric mean, as had been the case before 2010; and that the “goalposts” for the health, education, and standard of living indices had gone back to being stipulated a priori (e.g., 85 for life expectancy, 15 for average years of schooling, etc.) rather than being set, in accordance with the 2010 revisions, to the highest and lowest actual levels across all countries, causing the “goalposts” to shift every year (which is not helpful for time-series analysis). You can access the Human Development Report data directly by visiting this website.
UNICEF has published State of the World’s Children annually since 1980. Each edition includes statistical data on indicators related to the well-being of children. UNICEF’s Multiple Indicator Cluster Survey website has data on children and women, including updated infant and under-5 mortality statistics and access to recent survey reports and data.
The United Nations Population Division has a wide range of useful demographic data, some of which are accessible through this user-friendly data query system. The sex ratio by broad age groups (including the male-to-female ratio in the 0-4 age group, an important indicator of “missing girls”) is available here.
The Norwegian Social Science Data Services keeps current a useful MacroData Guide (in English) with links to state-of-the-art sources of cross-national data on demography, economics, education, health, labor and employment, crime, corruption, natural resources, politics, conflict, human rights, inequality, gender, religion, and other topics.
2. Health and Health Care Data
The World Health Organization’s Global Health Observatory (GHO) has a large cross-national data repository, especially for years since 1990, on mortality, disease incidence, child nutrition, child health, maternal and reproductive health, immunization, HIV/AIDS, tuberculosis, malaria, water and sanitation), non communicable diseases and risk factors, health systems, environmental health, injuries, and violence.
The Institute for Health Metrics and Evaluation (IHME) is a global health research center at the University of Washington, Seattle, USA. The IHME administers the Global Burden of Disease Project, which involves about 3,600 researchers in 145 countries measuring mortality and morbidity from 350 diseases and injuries in 195 countries from 1990 to the present. Global Burden of Disease and other health data collected by the IHME are available in The Global Health Data Exchange (GHDx), a data catalog created and supported by IHME. In January 2017, the IHME received a ten-year, $279 million grant from the Bill and Melinda Gates foundation to expand its data collection activity.
Since the mid-1980s, some 300 Demographic and Health Surveys (DHS) have been conducted in 90 developing countries. The surveys produce highly regarded data on maternal and child health service delivery and related indicators. The DHS website includes a useful stat complier.
3. Data for Latin America and the Caribbean
The Socio-Economic Database for Latin America and the Caribbean (SEDLAC), housed at the Center for Distributional, Labor and Social Studies (CEDLAS) at Argentina’s Universidad Nacional de La Plata, uses microdata from over 300 household surveys in 24 countries to produce comparable statistics on per capita income, income inequality, income poverty, household size, educational attainment, housing quality, durable goods ownership, access to electricity, safe water, and adequate sanitation, employment, and eligibility for disability and retirement pensions. Some of the variables are disaggregated by gender. The data pertain mostly to the period since 1990, but in some countries statistics go back as far as 1974. Country experts update the tables when microdata from a new survey become available, and new databases are published approximately twice per year. The website can be used in either English or Spanish.
The Pan American Health Organization (PAHO), an agency of the World Health Organization, provides health data on 48 countries and territories in the Western Hemisphere, from Canada to St. Lucia to Argentina. The site includes a statistical compiler called Core Indicators – Interactive Version.
4. Infant, Child, and Maternal Mortality
The infant mortality figures published in the World Bank’s World Development Indicators are based on estimates compiled by the Inter-agency Group on Child Mortality Estimation, which includes specialists from the World Bank, World Health Organization, UNICEF, and the United Nations Population Division. The census, survey, and vital registration data underlying these estimates are available at the child mortality website of the Interagency Group on Child Mortality Estimation. The Interagency Group has a similar site for maternal mortality data.
The method used to produce the Interagency Group infant and under-5 mortality estimates was described initially in Kenneth Hill et al., Trends in Child Mortality in the Developing World: 1960-1996. New York: UNICEF, 1999. “Levels and Trends of Child Mortality in 2006: Estimates Developed By the Inter-agency Group for Child Mortality Estimation” provides a more comprehensive account of the methodology employed, along with infant and under-5 mortality estimates for most of the world’s countries through 2005.
Since the mid-1980s, some 200 Demographic and Health Surveys (DHS) have been conducted in 75 developing countries. The surveys provide highly regarded data on infant and child mortality, among other indicators.
UNICEF’s Multiple Indicator Cluster Survey website has data on children and women, including updated infant and under-5 mortality statistics and access to recent survey reports and data. Designed and administered by UNICEF, other international organizations, and local government agencies, the MICS surveys are tailored to suit the particular informational needs of the host country.
You can query the United Nations Population Division’s World Population Prospects for infant and under-5 mortality estimates by country by five-year period (e.g., for 2010-2015). Data are available from 1950 to (worrisomely) 2100.
The UN Maternal Mortality Estimation Inter-Agency Group provides maternal mortality estimates for 171 countries in each year from 1985 to 2015. The data are available here and are described in a 2015 article in the journal Lancet.
5. Family Planning
The Demographic and Health Surveys (DHS) have data on indicators of family planning.
Researchers associated with the Track20 project of Avenir Health (formerly the Futures Group) have used an expert rating system to measure family planning program effort in 60 poor developing countries. The researchers send questionnaires to country experts, aggregate the responses into component measures of different aspects of family planning effort, and assign each country an overall score equal to its achieved percentage of the maximum attainable score on the combined components. Surveys were done in 1972, 1982, 1989, 1994, 1999, 2004, 2009, and 2014, and more frequently since then.
6. Vaccines and Immunizations
For Covid-19 vaccinations the OurWorldInData Covid-19 data explorer is reliable and extremely user-friendly.
In June 2000, researchers at UNICEF and the World Health Organization reviewed and evaluated all available immunization coverage information for as many countries as possible for as many years as possible from 1980 onward, seeking a “consensus estimate” of the share of a target population that had been immunized against a wide range of infectious diseases. To produce these estimates, they . The first estimates were released in 2001, Updated series are available at the World Health Organiation’s Global Health Observatory.
The World Health Organization (WHO) has data on infant and child undernourishment. Some countries have data spanning the early 1980s to the early 2000s. The sources of the data for each country are unusually well-described.
The Food and Agriculture Organization (FAO) has compiled data for about 190 countries on the proportion of the total, adult, and child populations that suffer from hunger (less than the calorie requirement for an active and healthy life in a particular country), as well as on many other food and nutrition-related indicators (food needs; food, protein, and micronutrient availability; food trade; food aid). The data are described in an appendix to the annual The State of Food Insecurity in the World.
8. Economic Affluence
The Penn World Table 10.0 has information on GDP in 183 countries for some or all of the years from 1950 to 2019. For studies comparing living standards across countries and over time, a particular GDP measure (RGDPe), expenditure-side real GDP at chained PPPs (in millions of 2017 US$), is recommended.
The Maddison Project Database in its 2020 release provides statistics on population and real GDP per capita for most countries of the world from 1900 (data are available for 45 countries) to 2018 (data are available for 169 countries). Some countries have data for earlier years (indeed, data are available for 14 countries in the year 1 C. E.) The figures are given for purchasing power parity and are presented in constant 2011 dollars. For many countries the data go back much farther than 1950: data are available for 43 countries in 1940, for 45 countries in 1900, for 36 countries in 1850, for 19 countries in 1800, for 5 countries in 1000, and for 14 countries in the year 1 C.E. Regional estimates (for Western Europe, Sub-Saharan Africa, Latin America, etc.) are available for 2016, 2017, and 2018, as well as for every 10th year from 1820 to 2010.
As of early 2017 the World Bank World Development Indicators included 17 measures of GDP (Gross Domestic Product) and 14 indicators of GNI (Gross National Income, formerly known as Gross National Product). The measures differ from one another on several dimensions, including (1) levels vs. growth rates, (2) total vs. per capita vs. per worker vs. per unit of energy use, (3) current US dollars vs. constant international dollars, and (4) market exchange rates vs. purchasing power parity. For studies comparing living standards across countries and over time, use GDP or GNI per capita figures in constant international dollars at purchasing power parity. As of early 2017 these variables were labeled “GDP per capita, PPP (constant 2011 international $)” and “GNI per capita, PPP (constant 2011 international $).” GDP is a measure of output; GNI is a measure of income. Among 149 countries with 2014 data on both of these indicators, the difference exceeded 10 percent in only 10 countries. In general, GNI is higher than GDP in countries that receive a lot of remittances from foreign workers (e.g., Philippines or Bangladesh) or that have (usually oil-funded) sovereign wealth funds that receive a lot of interest and dividends from investments in foreign countries (Timor Leste, Kuwait, Norway). Conversely, GDP is generally higher than GNI in countries that have high levels of foreign direct investment, often in oil or mineral extraction, and in which a large share of earnings flow back to the host country in the form of repatriated profits (Equatorial Guinea, Ireland, Mongolia, etc.). For the variables “GDP per capita, PPP (constant 2011 international $)” and “GNI per capita, PPP (constant 2011 international $),” data as of early 2017 were available only for the years 1990-2015. For earlier years, use the Penn World Table or Maddison Project data on GDP per capita in constant dollars at purchasing power parity.
9. Educational Attainment
Data on average years of schooling and other measures of educational attainment, as well as the same indicators for females only and for males only, are available at www.barrolee.com for 146 countries at five year intervals from 1950 to 2015. A particularly useful indicator of educational attainment is “average years of total schooling” in Column “L” of the databases you can download at this site.. If you have access to the Journal of Development Economics, you can also consult an article describing the methods used to produce the data: Barro, Robert J., and Jong Wha Lee. “A new data set of educational attainment in the world, 1950–2010.” Journal of Development Economics 104 (2013): 184-198. A “Long-term Data” section of the Barro and Lee website includes “estimated school enrollment ratios from 1820 to 2010 and estimated educational attainment for the total, female and male populations from 1870 to 2010. The estimates are available in five-year intervals for 111 countries.”
For illiteracy, see the World Bank World Development Indicators and the UNDP Human Development Reports listed at the top of this page under “Major Statistical Compendia.”
The UNESCO Institute for Statistics has recent data on enrollment ratios, repetition rates, and other educational indicators for most of the world’s countries.
10. Income Inequality
The World Inequality Database is the state of the art in providing high-quality data on income and wealth inequality, as well as on a variety of indicators of average and total production and income broken down in various ways. The database, which went public in 2015, has an exceptionally user-friendly web interface. Most previous estimates of income inequality have relied on survey data. The World Inequality Database combines survey data with national accounts statistics and income tax data. In most countries national accounts statistics and income tax data go back farther in time than survey data. The World Inequality Database methodology thus allows for for the production of a longer time series (especially in wealthier countries). Also, surveys notoriously underreport the incomes of the rich and the very rich; the other sources of data help to correct for this problem. Another distinctive advantage of this data source is that for some countries it provides information on wealth inequality as well as on income inequality (wealth is a stock, income is a flow).
A good source of national income inequality data for a large number of countries around the world is the World Income Inequality Database hosted by the World Institute of Development Economics Research (WIDER) in Helsinki. The May 6, 2020 version of the data set covers 200 countries with more than 11,000 observations from as recently as 2018. You’ll need the .pdf user guide as well as the Excel spreadsheet file.
The Socio-Economic Database for Latin America and the Caribbean (SEDLAC) based at the Universidad Nacional de La Plata in Argentina calculates income inequality measures and poverty headcounts directly from survey microdata in each of 25 countries in the region. Although data for the 1980s and earlier is sparse, SEDLAC is a good source for income inequality and income poverty data from Latin America from 1990 to the present.
A useful database of Gini coefficients (measures of income inequality) is Branco Milanovic’s All the Ginis. Gini coefficients range from 0 (everyone has the same amount of money) to 1 (one person has all the money, everyone else has none). You’ll need the description as well as the data. The link provides the data in Stata (.dta) format; if you would prefer Excel (.xlsx) I have converted the .dta file to .xlsx and stored it here. In cases where more than one Gini estimate exists for a given country-year, the “Giniall” column (Col. BD) indicates which one Milanovic finds most credible. The All the Ginis database does not use interpolation or extrapolation to fill in missing values, specifies the source from which each Gini was obtained, and indicates whether each Gini pertains to income or consumption (consumption Ginis tend to be lower) and, if it pertains to income, whether it registers the distribution of pre-tax-and-transfer or post-tax-and-transfer income (post-tax-and-transfer Ginis tend to be lower).
All the Ginis also indicates whether a particular Gini pertains to distribution across households or individuals. In the USA, the rich are less likely than the poor to be single, or single parents. So, rich households tend to be larger than poor households, which means that their higher incomes are divided among a larger number of people. This depresses the income share of rich individuals relative to rich households, and makes Ginis adjusted for household size (“equivalence adjusted”) lower than Ginis not so adjusted. In poor countries, rich families tend to have fewer children than poor families. So, rich households tend to be smaller than poor households, which means that their higher incomes are divided among a smaller number of individuals. This increases the income share of rich individuals relative to rich households, and makes Ginis adjusted for household size (“equivalence adjusted”) higher than Ginis not so adjusted.
The Global Consumption and Income Project (GCID), initiated in April 2016, provides annual Gini coefficients for 160 countries (Lahoti, Jayadev, and Reddy 2016). Its website has a data visualization application showing time series for up to four countries from 1960 to 2015…when it works, which was not always the case when I tried it on a couple of different browsers. A distinctive feature of this database is that it separates Gini estimates based on income from Gini estimates based on consumption. It uses interpolation and extrapolation to fill in missing values. As of February 2017 they appeared still to be working out the kinks…the income series for South Korea is not plausible.
The Standardized World Income Inequality Database version 5.1 (SWIID) developed by Frederick Solt provides annual Gini coefficients for 176 countries for as many years as possible from 1960 to 2014. Drawing on data collected by more than a dozen well-known compendia, it uses statistical techniques (including extrapolation and interpolation) to provide Ginis based on both gross and net per capita income, providing 95 percent confidence intervals around each annual point estimate. The website has a data visualization application. The data are described in Frederick Solt, “The Standardized World Income Inequality Database,” Social Science Quarterly 97.5 (2016), 1267-1281.
The strengths and weaknesses of the major compendia of data on income inequality are discussed in “Appraising Cross-National Income Inequality Databases,” a special issue of the Journal of Economic Inequality 13 No. 4 (December 2015).
11. Income Poverty
This webpage is the gateway to the World Bank’s data on income poverty and income inequality. Particularly useful is the World Bank’s PovcalNet database. It includes “estimates of global poverty from 1981 to 2013 based on 2011 PPP [purchasing power parity]. The new poverty estimates combine Purchasing Power Parity (PPP) exchange rates for household consumption from the 2011 International Comparison Program with data from more than one thousand five hundred household surveys across 163 countries in the world, including 25 other high income countries. Over two million randomly sampled households were interviewed for the 2013 estimate, representing 85 percent of the population of the whole world.”
Only a subset of countries have annual, or nearly-annual, data for the entire period from 1981 to 2013 (some countries have data from later years as well), but the PovcalNet database is probably the most complete and reliable compendium of income poverty data in existence. To get the poverty data for Brazil (for example):
- Go to the PovcalNet database home page
- Click “Choose your countries/aggregates” in the upper right-hand corner of the home page
- Select “Brazil” from the list of countries, click > , then click “Continue >>”
- If the default $1.90/day poverty line is OK with you (an alternative is $3.20, you can change it), and you want data for all available years, click “Select all” next to the heading “Year” in the box with all available years. Under “Output style” check the “flat table” box, then click “submit.”
- You will get a table with the years down the side (each year from 1981 to 2015 in which Brazil’s periodic PNAD survey was fielded) and variables across the top. The “headcount” variable is the share of the population living in households whose incomes are below the income poverty line (of $1.90 per day, if that is the line you set or accepted). The “poverty gap” is an index of the average severity of poverty; it measures how far the average poor person falls below the poverty line. Also useful are the “Mean ($/Month),” which gives the mean monthly per capita income in 2011 $PPP of all households surveyed, and “Median,” which gives the median monthly per capita income in 2011 $PPP of all households surveyed. Also available are the Gini index of income distribution based on the survey data and the country’s population in the survey year.
- NOTE: Brazil’s poverty data are based on reported household income. For many other countries, especially in Africa and Asia, poverty data are based on reported household consumption expenditure. For cross-national and over-time comparisons, income vs. consumption expenditure makes a big difference. Poverty headcounts and Gini indices of income inequality each tend to be lower when the poverty measure is based on consumption rather than income is , but this is not always the case. The safest course of action is never to compare a poverty headcount (or poverty gap) based on income to a poverty headcount (or poverty gap) based on consumption.
- To copy the Brazil poverty data into a word-processing, statistical, or spreadsheet program (e.g., to calculate percentage change or to create a graph), click “copy.” A small window appears with some html text, and above that window is “Ctrl-C to copy the content to clipboard.” The html text should be highlighted, so just type “Control-c” on your keyboard and the table will go into your clipboard. You can then paste it into (for example) a Microsoft Excel spreadsheet. The rows and columns will line up just as they do on the screen in your browser.
12. Water and Sanitation
The WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation provides carefully collected data on the proportion of the population with access to safe water and adequate sanitation. The most detailed information is to be found in the country files, some of which contain estimates from as far back as 1980. In most cases, however, UNICEF considers data collected before 1990 as significantly lower in quality than data from 1990 forward.
13. Geographical Variables
Data on land area, proportion of the population near the coast, latitude, population, and other such variables have been assembled by John Gallup, Andrew Mellinger, and Jeffrey Sachs. You can find data on these and other geographical variables in a section of the Dataverse at Harvard University (2010 version).
14. Democracy, Civil and Political Rights, Women in Parliament, Constitutions
Varieties of Democracy Project
State-of-the-art in data on democracy and civil and political rights, as well as on many related indicators, have been compiled by the Varieties of Democracy (V-Dem) project. V-Dem began data collection in 2012 and went public in 2016. Version 9 (May 2019) had 27 million observations coded by more than 3000 experts on 471 democracy-related indicators for 202 countries (V-Dem Institute 2019a). Multiple coders rated each indicator, and the V-Dem administrators used transparent methods to reconcile the ratings into country-year point estimates with uncertainty bands. In addition to hundreds of individual indicators related to democracy, V-Dem provides composite indices.
Among the most widely-used of V-Dem’s composite indices is the V-Dem Liberal Democracy Index, which consists of three terms weighted and summed:
(1) the V-Dem Electoral Democracy Index (weighted 25 percent, bundles five components: popular elections, electoral integrity, suffrage breadth, freedom of association, and freedom of expression).
(2) the V-Dem Liberal Component Index (weighted 25 percent, consisting of 14 questions pertaining to equality before the law and individual liberty and 9 questions pertaining to judicial and executive constraints on the executive).
(3) the product of the Electoral Democracy and Liberal Component indices (weighted 50 percent).
The V-Dem Liberal Democracy Index is a valid index of whether a country holds free, fair, and inclusive elections, protects basic rights, and allows the legislature and judiciary to constrain the executive. What the Liberal Democracy Index lacks is an indicator of the degree to which elections are decisive — whether election victors have meaningful power to rule. In the absence of such an indicator, military-fist-in-civilian-glove or other hybrid regimes could be misclassified as liberal democracies.
V-Dem, however, provides indicators that could be used to represent the degree to which elections are decisive. Experts are asked (1) whether, in formulating domestic policy, a head of state and/or head of government must seek approval from ruling parties, royal councils, the military, religious leaders, tribal or ethnic councils, or other unelected actors of comparable scope and power, (2) whether the state is autonomous from the control of other states in domestic and foreign policy, and (3) what percentage of the national territory is effectively controlled by the national state (V-Dem Institute 2019b). Such indicators could be aggregated into an “Electoral Decisiveness Index” and combined with the V-Dem Liberal Democracy Index into a “Full Democracy Index,” which would be a valid measure of democracy.
V-Dem uses an ordinal scale, like the Polity IV and Freedom House indicators discussed below. The Democracy-Dictatorship Index, also discussed below, is a dichotomy. For the terms democratic and authoritarian to be meaningful, there must indeed be a threshold below which a regime becomes authoritarian, regardless of whether a graded or dichotomous measure is used. Also, if scholars are interested in comparing democratic to authoritarian regimes, or in assessing the causes or consequences of the longevity of a democratic regime, some sort of categorical indicator is indispensable. For a rigorous effort to use V-Dem data to create a four-category regime classification (liberal democracy, electoral democracy, electoral autocracy, closed autocracy) see Anna Lührmann, Marcus Tannenberg, and Staffan I. Lindberg (2018). “Regimes of the World (RoW): Opening New Avenues for the Comparative Study of Political Regimes.” Politics & Governance 6 No. 1, 1-18.
Prior to V-Dem, the most comprehensive dataset of quantitative democracy indicators was probably Polity IV, compiled by Ted Robert Gurr, Keith Jaggers, Monty Marshall, and their collaborators. The Polity IV data go back to a country’s date of independence or to 1800, whichever is more recent. Polity IV has transparent and detailed coding rules, uses multiple coders for each country-year, and tests for inter-coder reliability. To create the database, coders drawing on secondary literature assigned each of the world’s independent nations, in each year from 1800 to 2018, scores on “democracy” and “autocracy.” The number of countries coded ranged from 22 in 1800 to 167 in 2018. The scores are based on three sets of criteria: (1) “openness and competitiveness of the recruitment of the chief executive”; (2) “constraints on the authority of the chief executive”; and (3) “political participation and opposition.” Each criterion has subcomponents. For example, political participation and opposition includes “regulation of participation” (how much factionalism and personalism there is in politics) and “competitiveness of participation” (how much incumbents restrict political opposition). The subcomponents and components are scored, weighted, and combined to form a democracy score ranging from 10 to 0, as well as an autocracy score ranging from 0 to 10 (10 is the rating assigned to the most democratic democracy and to the most autocratic autocracy). The two scores are sometimes combined to form a “polity2” score ranging from 10 (most democratic) to -10 (most autocratic). Unlike the Democracy score, the polity2 score implies a penalty for more autocratic autocracy, as well as a premium for more democratic democracy. The coding is done transparently and systematically and is checked for inter-coder consistency.
To obtain the data, download http://www.systemicpeace.org/inscr/p4v2018.xls You’ll also need the User’s Manual at http://www.systemicpeace.org/inscr/p4manualv2018.pdf The p4v2018.xls file has the country name in Column D, the year in Column E, the Democracy score in Column H, the Autocracy score in column I, and the polity2 score in Column K. Scholars doing quantitative analyses often add 10 to the polity2 score in order to make their research findings easier to interpret. Adding 10 makes the polity2 scores run from 0 to 20 instead of from -10 to +10.
Very useful to the assessment of democratic backsliding using the Polity IV data are the annual summaries of changes that the administrators made to country ratings (components, subcomponents, summary democracy and autocracy scores) in a particular year. Descriptions of every change made in a particular year, and of the reasons for such changes, are available for each year from 2012 to 2018 inclusive at http://www.systemicpeace.org/inscr/ Scroll down until you reach files with titles lile p4ch2012.xls, p4ch2013.xls, etc.
Freedom House since 1972 has rated countries annually on “political rights” and “civil liberties.” Go here and scroll down to “Freedom in the World Comparative and Historical Data.” One of the links under this heading allows you to download to your desktop an Excel spreadsheet titled “Country and Territory Ratings and Statuses,” which has the political rights and civil liberties time series for 195 countries and 15 territories from 1972 to present. The methodology by which the 2019 scores were produced is described here.
Freedom House since 2003 has rated countries annually on a finer 100-point scale, with political rights given 40 points (electoral process 12, political pluralism and participation 16, functioning of government 12) and civil liberties given 60 points (freedom of expression and belief 16, associational and organizational rights 12, rule of law 16, and personal autonomy and individual rights 16). These ratings (for the seven subcomponents, as well as in aggregate) are available here as “Aggregate Category and Subcategory Scores, 2003-2019 (Excel).” An even finer breakdown into 10 separate political rights scores and 15 separate civil liberties scores is available here for countries for the years 2013-2019 as “All Data FIW 2013-2019 (Excel).”
The Democracy-Dictatorship index introduced by Alvarez et al. (1996) and revised and extended by Cheibub et al. (2010) is based on a minimalist conception of democracy, and unlike the V-Dem, Freedom House, and Polity IV indices, which use ordinal scales, the Democracy-Dictatorship index is a dichotomy. If a regime has an executive and legislature chosen in elections with more than one party, and at least one alternation in power under the same electoral rules as brought the incumbent to office, it is a democracy. If not, it is a dictatorship (Alvarez et al. 1996: 18; Cheibub et al. 2010: 69). Cheibub et al. (2010: 78) defended the dichotomous character of the Democracy-Dictatorship index by arguing that the use of a graded (continuous or polychotomous) notion of democracy will inevitably lead to absurdities, compelling the observer “to speak of positive levels of democracy in places like…Chile under Pinochet or Brazil during the military dictatorship” (Alvarez et al. 1996: 21). Such a pitfall might be sidestepped, of course, simply by establishing a threshold below which the quality of democracy would be zero (Collier and Adcock 1999: 548-550). As noted above, however, for the terms democratic and authoritarian to be meaningful, there must indeed be a threshold below which a regime becomes authoritarian, regardless of whether a graded or dichotomous measure is used. Also, if scholars are interested in comparing democratic to authoritarian regimes, or in assessing the causes or consequences of the longevity of a democratic regime, some sort of categorical indicator is indispensable.
The dichotomous Democracy-Dictatorship index is nevertheless rather more generous in awarding the designation “democratic” than many alternative operationalizations. For one thing, a regime can earn the label “democratic” even if elections are unfair. Alvarez et al. (1996: 20) noted that “some fraud is a ubiquitous phenomenon in democracies… We concluded that there is no way to assess the validity of such allegations in a standardized way.” In most alternative interpretations, moreover, for a regime to be designated a democracy, almost all adult citizens should have the right to vote and to hold elective office. The Democracy-Dictatorship Index, however, excludes universal suffrage as a criterion for democracy, citing evidence from Western Europe and Latin America that “indicates that the distribution of votes across parties changes only slowly after each extension of suffrage, implying that even when suffrage is highly restricted, divergent interests are being represented” (Alvarez et al. 1996: 19). Apparently, then, for the Democracy-Dictatorship Index, the representation of divergent interests suffices to warrant the designation democratic.
Also explicitly excluded from the Democracy-Dictatorship decision is any measure of individual rights, on the grounds that “if democracy requires civil liberties, political rights, freedom of the press, and other freedoms, then inquiries about the connection between democracy and such freedoms are…precluded” (Cheibub et al. 2010: 73). It is far from self-evident, however, that treating electoral contestation and basic rights as elements of democracy would preclude questions about their connections. The Varieties of Democracy project, discussed above, provides separate indices of contested elections and of basic rights. It would not be hard to find out how closely the two indices are correlated with each other, but what would a high positive correlation mean? Would it mean that basic rights are causing contested elections, or that contested elections are causing basic rights, or that both are being caused by a third factor, among which a strong candidate would be a degree of functional interdependence sufficient to characterize basic rights and contested elections as parts of a single whole — democracy?
The Democracy-Dictatorship data cover 202 countries from 1946 or year of independence to 2008. The dataset, codebook, and associated research paper are available here.
Comparisons and Critiques of Quantitative Democracy-Rating Projects
Descriptions and critiques of quantitative democracy indicators, including the Polity and Freedom House indicators,
Coppedge, Michael, et al. “V-Dem comparisons and contrasts with other measurement projects.” V-Dem Working Paper 45 (2017).
Doorenspleet, Renske (2018). “The Numerical Value of Democracy: League Tables, Scores and Trends.” Chapter 2 (29-68) in Renske Doorenspleet, Rethinking the Value of Democracy: A Comparative Perspective. Cham, Switzerland: Palgrave McMillan.
Munck, Gerardo, and Jay Verkuilen, “Conceptualizing and Measuring Democracy: Evaluating Alternative Indices.” Comparative Political Studies 35 No. 1 (February 2002), 5-34. The article and commentary on it are available here if you or your institution subscribe to this journal.
Gender Quotas, Voter Turnout, Electoral Systems, etc.
The International Institute for Democracy and Electoral Assistance (IDEA) has useful databases on voter turnout, electoral systems, gender quotas for national legislative seats, and other indicators related to elections and legislatures. Click here to access the databases.
The Constitutions Project provides data on the powers constitutionally assigned to the executive, legislative, and judicial branches of government, as well as the text of each of the world’s constitutions. The interface is exceptionally user-friendly, with lots of graphics.
15. State Capacity
A World Bank webpage provides access to aggregate governance indicators for 212 countries for 1996-2007 for six dimensions of governance: voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption.
16. Free-Market Orientation
The Fraser Institute in Vancouver, BC, rates most countries of the world according to how closely each conforms to what the Institute defines as a free-market system. The data are at the Fraser Institute website.