Thread by @bbherre

🚨New Dataset Alert!🚨

What economic ideologies do political leaders have around the world?

A 🧵 on what we learn from the Global Leader Ideology dataset (GLI) I have built over the last four years, and why it matters.

#polisciresearch #socsciresearch

Social scientists have long studied — and often found — that leftist governments pursue different policies than rightist governments (read a recent summary here: bit.ly/3yNHzVn). But this research has mainly focused on OECD countries, neglecting other parts of the world.

This focus on OECD countries in part is because researchers often use off-the-shelf datasets which only provide data on the ideologies of leaders and parties in OECD countries. This is where my GLI dataset comes in.

The dataset vastly expands the scope of previous datasets and refines their coding by classifying political leaders as leftist, centrist, rightist, or non-ideological in 182 countries annually from 1945 (or independence) to 2020.

So, what do these data show? For a guided tour, continue with this thread. If you would rather explore the data yourself, you can download it from GitHub (bit.ly/3H69bry) or from my homepage: bastianherre.com/data

While the following graphs look at political leaders, understood as the most politically powerful person (as identified by the Archigos dataset, @hgoemans et al.), the dataset also covers heads of government.

The data show that most political leaders around the world have identifiable ideologies: we were able to identify a leader’s ideology for more than 95% of country-years. Across all countries, years with leftist leaders are slightly more common than years with rightist leaders.

The data also show that the frequency of ideologies differs between democracies and dictatorships (as coded per @vdeminstitute). In democracies, leftist and rightist leaders are equally common. In dictatorships, leftist leaders are clearly more common than rightist leaders.

Looking at how leader ideologies differ across countries, I find that some countries have long been governed by leftist or rightist leaders (those in darker shades). But many countries have indeed been led by both leftists and rightists in the last decades (in lighter shades).

In Western Europe and North America, rightist leaders have been especially common; but they often alternated in power with leftists. Centrist leaders were mostly concentrated in a few countries. But what about the political leaders beyond these often-studied countries?

Eastern Europe and Central Asia were led by leftists until the end of the Cold War. They held on to power in some countries, and have regained power in others; but rightist leaders have also become predominant in some countries.

The Middle East and North Africa have seen long spells of both leftist leaders (especially in the Mediterranean) and rightist leaders (especially on the Arabian Peninsula). Though for a fair number of others, we were unable to identify an economic ideology.

In Sub-Saharan Africa, leftist leaders have been widespread; especially in Eastern and Southern Africa, and across the region in recent years; though rightist leaders also managed to stay in power for a long time in a fair number of countries.

In Asia and Pacific, some countries have been primarily led by leftists (particularly in South and East Asia) and others especially by rightists (particularly in Southeast Asia). In recent decades, power has changed more often between leaders with different ideologies.

And in Latin America and the Caribbean, most countries have been governed by both leftist and rightist leaders. While rightist leaders predominated in the past, in recent decades leftists have held power more frequently.

When looking at trends in leader ideologies over time, the data show that leftist leaders became increasingly common in the first few decades after World War II, until they became the most common group in the 1970s and 1980s. They slightly lost ground in recent decades.

Finally, the data demonstrate different trends between democracies and dictatorships: Leftist democratic leaders were only predominant in the early 1970s and have fluctuated since, while in dictatorships leftist leaders have been predominant since the 1970s.

Overall, the data highlight that political systems around the globe do not only differ in their political institutions, but also in the actors that run them.

In addition to the information on their ideologies, the dataset identifies the parties of political leaders, and includes variables — such as PartyFacts IDs — which make it easy to link the data to related datasets on political leaders and parties.

If you want to learn more about specific countries and leaders, my RAs and I have written hundreds of pages of country profiles which summarize our sources. Some of the profiles are a bit messy, but I hope what they lack in polish, they more than make up for it in transparency.

If you want to learn more about the dataset in general, I have written a working paper that describes the dataset’s contents and coding in detail, compares it to existing datasets, and illustrates its uses a bit more: bastianherre.com/research

I hope that the dataset will be helpful to anyone interested in studying why leaders differ in their ideologies worldwide, and to those wanting to study how leader ideologies affect policymaking and social welfare around the globe.

Building this dataset took a village. I am grateful to the many, many people who have supported me along the way. My special thanks go to my many research assistants, including @the_christhilf, @JuliaLodoen, @lysmavridis, and @ivanna_shevel.

I am also thankful to my dissertation committee, including @SlaterPolitics and @mikealbertus, to @g_arteagoitia, @carlhknutsen, @tobirommel and @torewig for their comments on the project, and my new employer @OurWorldinData.

I further thank all the creators of the existing datasets on which I partially rely. This includes the creators of the Comparative Political Dataset (by @DWeisstanner, @sa_engler et al.), the Database of Political Institutions (by @cesicruz et al.),

the Global Party Survey (by @PippaN15), V-Party (@AnnaLuehrmann, @StaffanILindber, @M_higashijima, @kailmarkvart, @a2hicken, @melislaebens, @medzihorsky, @AnjaNeundorf, @reutertweets, @keithweghorst et many al.), CHISOLS (by @BAshleyLeeds et al.),

, REIGN (@curtismbell, @ClaytonBesaw et al.), as well as Thomas Brambor et al., Ben Cahoon, Evelyne Huber and John Stephens, Dulce Manzano, the Perspective Monde team, and B. Schemmel for their respective datasets and projects on leaders’ ideologies and parties,

and @xuyiqing and @liulch16 for their panelview R package I used to create the region graphs. Finally, this project would also not have been possible without substantial financial support from the project’s funders, including @PearsonInst, @TheIHS, @StiglerCenter, @UChicagoCISSR.

All remaining errors in the dataset and the country profiles are my own. If you spot any, or have additional specific sources I missed, please send me a message here or email me at [e-mail]. I look forward to hearing from you!

If you have made it this far and think this dataset could be interesting to others: like this thread, retweet it, or tell them about the data! Thanks a lot!

And if you would like to explore the data further now, here are the links to my homepage (bastianherre.com/data) and GitHub again: bit.ly/3H69bry