The British Council and the University of Edinburgh have teamed up to prepare an empirical analysis of the actual effects of soft power (my introduction on soft power and the arts is here). How does this work?
The idea is to take measures of different assets of a country that we think might contribute to its soft power, and to take measures of outcomes that we think might be influenced by soft power, to see if there’s any actual evidence of impact. The report has a lengthy, careful review of what we know and don’t know about soft power, the distinction with ‘cultural diplomacy’, and clear presentation of data and how the results were obtained. The aim is to see whether soft power actually affects the choices made by people from other countries, and not simply whether soft power induces warm feelings about the place without it actually leading to anything (although Andrew Rose found that warm feelings generate greater exports, all else equal).
The modeling here is reduced form; i.e., there is no theoretical scaffolding that would generate the equations that are actually estimated. Instead, the researchers choose a short list of what could be considered soft power assets, and a short list of desired outcomes, and estimate whether the former seem to “cause” the latter. It’s a common technique, and about the only option in a study like this, since there does not exist any formal model of what soft power is or how it might work. But it does mean that the variables chosen are a judgment call. Different researchers might have chosen different variables, but they would still have to make a judgment call too. All this means is that we need to be cautious about reading too much into the results of this particular study; it will be the larger empirical literature on the topic, working with many different variables and with different results, that begins to generate a picture of what might be going on (as an aside: this ought to be the rule for all new social science papers – forget the headlines of “a new study shows…” and focus on what the field as a whole has to say).
In this paper, the soft power assets (in the regressions, the right-hand side variables) are; a measure of political democracy and freedom; per capita income; and three measures of cultural assets, namely the international reach of cultural institutions (a data series developed specifically for this study by the authors), a country’s cultural ranking in the “Good Country Index”, and the proportion of the population connected to the internet. The outcomes of soft power (in the regressions, the left-hand side variables) are four in number: incoming international students; incoming international tourists; levels of foreign direct investment (FDI) and whether votes in the United Nations are drawn towards the soft power. Data are for average values from 2000-2012.
What jumps out is the question of missing variables. If we take international students for example, the reputation of the country’s universities is going to have obvious importance – the US and the UK have advantages there that go beyond any of the soft power assets included in this model. The English language will matter, the infrastructure for accepting foreign students, immigration rules on staying in-country after obtaining a degree and so on and so on. FDI might be affected by these variables, but also be driven by closeness to markets, avoiding tariff barriers, low costs, etc. I’m not sure Mercedes-Benz opened its major manufacturing plant in Alabama because of the soft power of the state (though, to be fair, locating in Tuscaloosa County nearby the U of A may well be based on cultural as well as educational considerations).
So, when the regressions find “statistically significant” correlations between assets and outcomes, I’m not entirely surprised, but cannot help but feel we are missing a lot of the story, and must be very careful in making inferences. The press release makes the claim:
Cultural institutions, like the British Council and Goethe-Instuit, were found to be influential for attracting international students, international tourists, and FDI. The more countries that host a cultural institute, the better the return for the parent state. For example, a 1% increase in the number of countries a cultural institution from country X covers results in almost 0.66 percent increase in FDI for that country. In 2016 such a rise would have been worth £1.3bn for the UK, which recorded £197bn of foreign investment. It also prompts a 0.73% increase in international students for its country of origin. Using the latest UK figures from 2015/16, this equates to almost 3,200 additional international students.
And, well, I just cannot have confidence in the claim that if any country made an effort to increase by 1% its international ties by cultural institutions it would lead to an expected 0.66% increase in FDI. Yes, that’s the regression estimate, but the model is too crude to give us that sort of confidence. That’s not a criticism of the modeling: with a limited number of countries and years, the model has to be crude – you cannot work with 100 right-hand side variables. But it does mean that one has to be very careful in making claims. The British Council is asking us to read more into the estimated parameters than is really there.
So, let’s have more studies like this, that approach the question of soft power empirically, but from different angles. Let a thousand flowers bloom, so to speak, and see what starts to arise.
Next post: the King’s College soft power study.
After the war, the USA established “America Houses” in most major German cities to represent American culture. They were very popular and much appreciated, but by the 80s, most had been eliminated. The UK maintained the presence of its British Councils, which continue to be much admired and appreciated. Even though many factors shape impressions Germans have of the USA and UK, one can see the contrasting effects of the ongoing British Councils have compared to the loss of influence created by the elimination of the America Houses. This influence (or its loss in the US case) is too ephemeral and complex to measure, but it is readily apparent.