A Strange Study on Italian Nepotism


These are dark days for Italy. The country’s bond yields are way up; Prime Minister Silvio Berlusconi looks to be on his way out. And Italian soccer superstar Antonio Cassano is in the hospital recovering from a suspected stroke.

What better time then to blog about a strange new study about Italian nepotism? Authors Ruben Durante, Giovanna Labartino and Roberto Perotti study the effects that a 1998 law decentralizing the hiring process at Italian universities had on levels of nepotism. Pre-1998, candidates for academic positions were selected through a national process. After 1998, however, universities were given the power to hire their own professors. The researchers found that this decentralization led to increased nepotism in areas of “low civic capital,” but not in areas of “high civic capital.”  From the abstract:

Decentralization can lead to “good” or “bad” outcomes depending on the socio-cultural norms of the targeted communities. We investigate this issue by looking at the evolution of familism and nepotism in the Italian academia before and after the 1998 reform…

By far the most interesting part of this study is the researchers’ treatment of the term “civic capital,” something they define loosely, yet measure very narrowly. Their definition of a region with high civic capital is “an area where citizens are generally more politically involved and better informed,” and where individuals are “prone to internalize the social costs of their actions and the public is equally more likely to monitor the conduct of public officials.”

How do they measure such behavior? By just two things: the size of non-sport newspaper readership, and the rates of blood donation. So, what they’re essentially saying is that reading the news (outside of the sports page) and donating blood are strong indicators of one’s high civic capital. It’s certainly an interesting way of measuring a pretty vague concept. But surely there must be a more robust method. What about rates of voting? Or crime maybe?

Go ahead readers, how would you measure a region’s civic capital?



If I were italian, I'd measure a region's civic capital by hiring my brother to do so.


Probably much more difficult to quantify, but a measure of volunteership would be good.

-Percentage of parents in local schools that are members of PTA?
-Size/nature of coaches for children's sports leagues (i.e. for little league, one Dad who never played T-ball, or a staff of 5 who all played in college).
-Size/activity of volunteer fire department?
-Concur w/ voter registration & voter activity.
-Size/participation in neighborhood watch programs.

Civic capitol is an interesting concept, and I have a general sense that it deteriorates when Government begins providing a service that was previously left to the community.


Sometimes a nonsensical variable, like "civic capital", obtains prominence in a study only when it's realized it is mathematically needed to come up with an interpretation of the data that meets criteria of statistical significance.

But assuming they, a priori, always intended to use this variable, I would challenge that decision /hypothesis! It just doesn't make sense without further teasing apart. The "absence of crime " rating of a region seems a part of "Civic capital " more relevant than volunteerism for example.


Crime rates after 10 pm.


The authors are too narrow, but their methodology is fine. See Robert Putnam's "Bowling Alone" for a persuasive cases of how to measure "social capital". The idea is build an index of, ideally, dozens of theses indicators and combined them do overcome the methodological biases and shortcomings of any one measure. Principal Components Analysis can then do wonders.

In the US context: church attendance, bowling league or softball participation rates, tithing levels, sales data for playing cards, non-political club membership (et PTA, Elks Club, ...), blood donations rates (an unusually good marker, if I recall), rates of volunteerism, charitable giving rates, but most of all: church attendance. Crime rates are usually "explained" rather than "explanatory" variables, but that is methodological. A correlation is a correlation.

Not being an Italian sociologist, I don't have reasonable guesses for Italians. Tithing measures might be a good place to start.



Putnam argues the key to social capital is "Thin trust". "Thick trust" is trust amongst your relations and close friends, while "thin trust" is trust in people barely known to you. It takes a lot of social capital to make a joint stock company possible, for example.

A fantastic measure is to conduct a careful opinion poll around the question, "do you trust the others?”. Another measure is how"safe" do people think it is for a woman or child to walk about unescorted, especially at night? So, for example, Night clubs accessed by public transit might be evidence high levels of social capital.

In the Italian context, how do Bocce Ball rates or other "non-commercial" sports vary?

Another marker in the Italian context might be social acceptance of extra-judicial violence. For example, ask the question,"if a member of your family is raped by someone known to you family, is it better to report incident to the police or for your family to respond directly?" As one moves south from Naples, it's a bit murky if the Polizia or the 'Ndrangheta are the correct authority to use for framing.

Since social capital has been shown to have a positive effect on public health, perhaps you could work backwards from public health data (eg infant mortality rates) to social capital estimates.



BTW, I don't think the authors invented this measure of social capital, but instead adapted it from this paper:


Since those are U Chicago profs, why don't you walk down the hall and ask them?


Basically, the study found a centralized system has a level of nepotism that is the average of each region. Useful study.


It seems this method is.. as they say, "Abbastanza buono per lavoro di governo"

(Translated: good enough for government work)


Any variable that supports my world view.


Did anyone actually looked at the paper? I think it is a useful work in that it demonstrates (once more maybe) that decentralization per se is not a panacea, but that you have to be careful who you give power to. By the way, most of the alternative measures suggested in previous comments are either not applicable to Italy (there are no volunteers in fire departments there!) or are good but no data is available at the geographic level the authors look at (the province, really small if you look at a map of Italy). Regarding the measures of civic capital they use, sure they are not perfect, but, as Tylerh mentioned, the authors just follow the approach used in two of the most cited studies in this literature - by Putnam (newspaper) and Zingales & co. (blood donation) - so the same criticism should apply to them! You may want to tell your colleagues then and let us know what they say!


I would only point out that blood donation in Italy is completely free, i.e. you don't get paid for the blood you donate (they just offer you an italian breakfast).


I'm not sure if this would test CIVIC capital specifically, but I've thought of a (rather expensive) quantitative way to measure something like social capital or conscientiousness.

Researchers would leave a cheap wallet in some public place, perhaps with a €10 note inside and some basic information like a "if you find this wallet, please return to" note including a phone number or address. This is repeated in multiple regions, perhaps with variations that always include a small amount of money, some indication of ownership, and a route to return this money back to the rightful owner.

Researchers could see how often the wallet is returned, with or without the €10, and compare return rates from place to place. It is, though, an expensive project if return rates turn out to be low.


i LOVE it.
makes a ton of sense to me...all the good people I know rarely care about sports (publicly) and say they always give blood ;-)