Mapping the international mobility of political scientists

Pippa Norris was kind enough to share an advance version of the dataset she collected on academic political scientists around the world, which will be presented for the 50th anniversary of the European Consortium for Political Research. The datasets contains lots of variables about the education, place of work and career of political scientists employed at universities around the world. Because it includes information on the country where political scientists were born, where they got their degrees and where they work at the time of the survey, it is also suitable to visualise the international mobility of political scientists.

International mobility of political scientists

The first graph shows a network of political scientists who work in another country as the one where they were born: current country is the target, and country of birth is the source. I have restricted the data to individuals who hold a PhD, which results in a network of 445 individuals (the full dataset, which of course includes individuals who work in their country of birth, has 2446 entries). This variable is not able to differentiate whether people have grown up/studied in the country where they were born or not (we can gain some more information with the country of PhD as the source below). Moreover, the representativeness of this sample is not clear because of relatively low response rates for some countries, so we should be careful about generalisations.

Some interesting trends which correspond to intuition emerge, however. The United Kingdom appears as a central node in the network, mostly as a receiver of academics born elsewhere. The biggest exporters of political scientists to the UK are the United States and Germany, which has a negative migration saldo possibly due to the structure of the German academic job market (however, as mentioned below, the absence of academics with a German PhD working in the UK in the sample leaves a few doubts about the representativeness of the sample). The United States sends much more political scientists to Britain than the other way round; in contrast, the migration saldo between the UK and Australia is negative. Germany also sends a significant number of political scientists to the Netherlands, Switzerland and Scandinavia. What also comes out of the network is the centrality of Europe, North America and Australia in the mobility network.

Inflows and outflows of political scientists (country of birth-country of employment)

The data presented above uses country of birth as the origin of individuals. A different picture emerges if we take the country where they did their PhD as the origin, and the current country of employment as the target. This is what the second graph below shows. The nodes here are sized by out-degree: they are proportional to the number of outgoing PhDs . We can clearly see the dominance of the UK and the US in the network as the largest suppliers of PhDs for export. If we related this to the pattern presented in the previous network, this seems to involve essentially non-Brits obtaining their PhD in Britain and going elsewhere to work, possibly their country of birth. Italy occupies an interesting position in the network as a major exporter of PhDs, which can probably be accounted for by the European University Institute. There are some puzzling things in the data, however: strangely, there are only 3 political scientists in the dataset with a German PhD working in Britain, and none with a Dutch PhD. As mentioned before, the data should be taken with a pinch of salt; Ideally, this type of data could be collected from websites for completeness, as relying on surveys with variable response rates is bound to have a few problems.

Mobility of political scientists by where they got their PhD

Um mapa das relações familiares no Governo de Antonio Costa

Um mapa das relações familiares no Governo de Antonio Costa

Nas ultimas semanas, os jornais relataram um numero impressionante de relações familiares no governo de António Costa: casais, primos, irmãos a assumir cargos públicos em vários ministérios e secretariados de estado. Este artigo no Jornal Economico aponto a existência de uma rede de 50 pessoas e 20 famílias, todas elas ligadas ao governo e ao partido socialista.

Mesmo se o artigo do Jornal Economico é muito detalhado, é difícil ter uma vista global de esta rede. Isto é matéria ideal para uma análise de rede social (SNA). O mapa acima mostra a
rede de relações familiares no governo de António Costa usando o artigo do Jornal Economico. O que fiz concretamente: fiz uma lista de indivíduos e entidades governativas mencionadas no artigo e o tipo de ligações entre elas. As ligações familiares (e em certos casos amigáveis) estão indicadas em vermelho. Os indivíduos estão em vermelho e as entidades em cinzento. E possível aceder uma versão alta resolução se clicar na imagem.

Os dados utilizados para construir esta rede estão aqui. E possível eu ter esquecido algumas ligações, por isso é possível comentar na lista para adicionar indivíduos e entidades.

É importante clarificar que todas as relações no mapa não são necessariamente ilegais ou mesmo relacionadas com nepotismo. Acontece que quando se está envolvido na política, há uma grande probabilidade de este fato estruturar o resto das relações sociais (incluindo a escolha de cônjuge). Também é normal pessoas das mesmas famílias partilharem as mesmas ideias políticas. O que se torna problemático (para a democracia) é se ter ligações familiares torna-se uma condição para aceder certos cargos públicos. Finalmente, este mapa não implica que não há práticas semelhantes em outros partidos que estiveram no poder ( ver aqui.

Mapping Preferences over Brexit in the House of Commons

In the graph above (zoomable version here), I have mapped the votes of British MPs in the 8 options given to them on March 27 (a couple of hours ago) in indicative votes. This is a 2-mode network linking MPs and options for Brexit. The network graph shows how different options are related to each other: they are closer if a greater number of MPs have voted for them. Their size shows the number of favourable votes for each option (none of them obtained a majority. The graph show the high level of polarisation over these different options, with two clear poles: the “Hard Brexit” pole with a number of MPs for whom the only option is No Deal, or a preferential trade arrangement, and the Soft Brexit-No Brexit pole, linking the Customs Union, Labour’s Plan and a Second Referendum/Revocation of article 50. The EFTA/EEA and Common Market 2.0 options link the two poles, but there seems to be very few MPS that could be rallied in the middle.

Visualising the world’s 20 richest countries, 1800-2016 in one minute.

I made this animation using John Burn-Murdoch’s code and data from the Maddison Project Database.

Academic hierarchies in US political science

In a number of recent projects (see e.g here), I have been interested in the structure of the academic job market, how it is organised in different countries, and how it may shape the ideas that it produces (see here for a comparison of economists in the US and Germany). Something that we highlight in the latter paper is how concentrated and hierarchical the field of US economists is compared to Europe. Most top US economists are affiliated with a pretty small number of elite universities (Harvard, MIT, etc.). Is it different for other disciplines, for instance in political science? My hunch was that it isn’t.

I have just come across this paper by Kim and Grofman published in January 2019 on the “Political Science 400”, looking at the top-cited political scientists currently employed in US universities. The online appendix has some interesting data on these 400 top US political scientists, especially where they work now and where (and when) they got their PhD. This can give some interesting insights into the structure and hierarchies in US political science.

In the graph below, I have used this data to build a network of universities where each arrow represents a flow of political scientists in the top 400 linking the institution where they obtained their PhD (the source) and the institution where they worked in 2017 (the target). The size of the nodes represents the out-degree (the number of individuals who got their PhD from that institution) and the size of the arrows represents the number of academics

Network of US political scientists, 2017

What does the data show? First, we can see a “core” of prestigious universities who are fairly closely connected with each other: Harvard, Stanford, Berkeley, Princeton, Columbia, University of Chigago, Yale. In fact the top-5 granting institutions alone (Harvard, Berkeley, Michigan, Yale and Stanford) account for 40% of all the political scientists in the top-400. Harvard alone supplies 49 of the 400 top political scientists. The structures that comes out is indeed hierarchical: the “inner circle” of universities at the top mostly recruit from within the top-circle, other universities outside the circle recruit from the top, but there is fairly little “upward” mobility: few top universities recruit people from outside the top circle. Finally, access to the upper tier of US political science (and probably the whole field) is not very open to academics with a foreign PhD: only 22 (5.5%) of the top-400 have a PhD from a foreign institution.

Assigning podcasts as assignments to students

In my course on Global Challenges, I ask students to submit a policy podcast. Here are the guidelines that I give to students, and an example of what students have done.

Purpose of the assignment

The purpose of this assignment is for you to demonstrate the ability to use the theoretical tools provided in the course by a) analysing a concrete global challenge, and b) formulating policy proposals drawing on the theories discussed in the module. We ask you to put together a policy proposal podcast dealing with one out of three policy problems. You are expected to carry out some level of research on the problem at hand and collect the relevant data, analyse the causes of the problem, and formulate policy proposals to deal with the problem, by drawing on the lectures, the seminar discussions, and the readings of the module. The policy proposals should be grounded in the theories discussed in the module, but the language should be directed to a lay, informed audience.  

Three policy problems

You need to answer one of the following three questions:

1. Recently, the Italian government had decided to refuse entry to boats crossing the Mediterranean with migrants. How could we imagine a system where migrants seeking asylum are allocated fairly across European countries, in a way that is accepted by local citizens?

2. Donald Trump has announced that the United States would withdraw from the Paris climate accords signed in 2015. What kind of mechanism should we adopt so that large economies accept to reduce their carbon emissions?

3. How should we address the backlash against democracy and democratic norms in many parts of the world, including Europe? 


You are asked to submit 1 audio file and the script of your audio.

Your audio podcast should contain the following sections:

(1) an introduction presenting the problem

(2) an analysis of the problem which makes use of frameworks discussed in the module,

 (3) a policy recommendation proposing concrete measures to solve this problem, and a discussion of the main alternatives,

(4) a brief analysis of the pros/cons of the measure proposed

(5) a conclusion/opening

The podcast should be last no more than 10 minutes. This should correspond to ca. 1800 words of script (

Data and evidence

Someone reading your script should know your sources of data (literature or people) and the alternatives you considered as well as your final recommendation. However, you do not need to reference explicitly in spoken word. The reader should also learn enough about how you did your work to make an informed judgment on how seriously to take its recommendations. Ask yourself “Is this credible?” “Why would they take my advice seriously?” This also relates to the use of sources: make sure that you use reputable sources. While there is a wealth of information on blogs and websites, a policy recommendation relying only on this type of sources may not appear totally reliable. 


The podcast should be written in the style of a real piece of policy advice designed to influence a (busy) policy-maker, so it should be clear, concrete and to the point, but without jargon.

Assessment criteria

The assignments will be assessed by drawing on the elements in the table below. Use concise, clear language. Note that writing for the document should be readable, so please keep margins and font sizes reasonable. Make sure that the document has a narrative flow: the policy analysis tells a story, and should not just be a shopping list of points under a series of headings. This should be something that a policy-maker can read and understand fairly easily, and which will aid him/her in the decision-making process. In the real world, overly dense, jargon-filled, poorly researched, sloppily written and badly presented documents would be handed right back to you to be fixed (at best).

Assessment criteria

  1. Clarity, coherence and feasibility of policy proposal
  2. Link to concepts & issues handled in class
  3. Creativity (effort to make podcast interesting, fun, entertaining; evidence of creative thinking)
  4. Format & Quality (delivery)

A note on the format

You are allowed to use music or effects in the podcast, but you do not have to. You will not be assessed on the technical aspects of the podcast as such (audio quality, etc), but making it exciting and interesting will enhance the appeal of your policy proposal.

Resources for making podcasts & examples

Migrant Workers or Working Women? Comparing Labour Supply Policies in Post-War Europe

Forthcoming in Journal of Comparative Policy Analysis (access article)


Why did some European countries choose migrant labour to expand their labour force in the decades that followed World War II, while others opted for measures to expand female employment via welfare expansion? The paper argues that gender norms and left power resources were important structuring factors in these choices. Female employment required a substantial expansion of state intervention (e.g childcare; paid maternity leave). Meanwhile, migrant recruitment required minimal public investments, at least in the short term, and preserved traditional gender roles. Using the contrasting cases of Sweden and Switzerland, I argue that the combination of a weak left (labour unions and social-democratic parties) and conservative gender norms fostered the massive expansion of foreign labour and a late development of female labour force participation in Switzerland. In contrast, more progressive gender norms and a strong labour movement put an early end to guest worker programs in Sweden, and paved the way for policies to promote female labour force participation.

Keywords: labour migration; female employment; Sweden; Switzerland; comparative public policy