The Italian Referendum: 53.75% NO


The upcoming “Italian Referendum”, coined Renzirendum in the German space, which is to take place on December 4, calls Italian voters to decide on drastic and for many even inappropriate institutional reforms which would change the distribution of power among the Parliament, the State, the Regions and the country’s various administrative entities (see Wikipedia). The argument against the reforms is that the Italian (much like the German) political system was built to favor gridlock so as to avert the emergence of fascism. If gridlock is lifted it might be used by leaders who are less benevolent than Matteo Renzi…

The Italian prime, who is widely seen as Italy’s last hope and certainly as the last political dike holding the flood of populism of the likes of Beppe Grillo and its Five Star Movement, has threatened that an outcome of No in the referendum will imply his resignation. Which is why this is seen by many as a fateful point in time for the fourth largest economy of the Eurozone. What complicates matters more is that while a Renzi resignation will most likely imply political instability at a time where the Italian GDP per capita is flailing and its employment rate is well below the EU-28 average despite Renzi’s labor reforms, his reforms on the other hand, not only might be the wrong medicine for the wrong ailment they might even of different use in the hands of someone like Beppe Grillo whose Five Star Movement is in close proximity to capturing power and has totally different plans. Put all this drama in the context of the Brexit and the Donald Trump presidency and you have the makings of uncharted uncertainty. Which is what caused the 10 year Italian Bond yield to more than double since August. Which also makes forecasting this referendum interesting.

As of this writing the last known poll had one quarter undecideds with 53.1% of those who had decided having done so for NO and 46.9% for YES.

In Google we see it similarly 53.75% for NO. Here is how. The searches we looked are votare no -si and votare si -no. The last seven days these searches look as follows:

A couple of remarks are due. I would have liked to look for “votare no” -si and “votare si” -no. The combination of  quotes and exclusion seem however to no longer work in Google Trends. We thus get a lot of noise from searches of the type come votare tu si que vales which are related to what appears to be a TV talent show named Tu Si Que Vales. This is why we subtract the word vales.

The volume of our searches are tiny so the validity of any inference might be questionable. Comparing our searches with Serie A and porn teaches us to be humble about this method (but also about people and even democracy itself…):

A final reason to take this prediction with a grain of salt is of course that not only we do not know all the relevant searches (possibly ones with the word referendum are relevant but we have no parsimonious way of capturing them) but also while we can see that our top related searches are relevant to the referendum (after we take the TV show out) they may not necessarily imply voter intend in the desired direction. So our ability to identify our queries is only known to be half right.

1 thought on “The Italian Referendum: 53.75% NO”

Comments are closed.