The story describes what happened when Professor Spinellis (Athens University) was hired to develop a data mining solution, that it was hoped, would improve tax collection and what happened when he tried to have his resulting improved data used to collect actual taxes. But if you’re short on time, I’ll just tell you here what happened, reportedly, nothing.

The report begins by pointing out that the Greek government estimates that one third of taxes owed are never paid. That sounds like a pretty big problem! But a problem like that makes for pretty easy data mining so Spinellis was able to provide Greek regional tax offices with targets of tax-dodgers. Unfortunately no increase in revenue followed his sharing of the improved targeting.

Spinellis then shifted his attention to a more local level where he looked at reporting success metrics in terms of cases closed to see if it was possible, through greater visibility, to improve performance at the local offices. This had little effect.

Apparently about two years into his efforts Spinellis realized that (emphasis added):

Fixing Greece’s tax system, and ultimately making the Greek economy work, was not a matter of tweaking his computer programs. It was not an information problem. It was a culture problem.

At the same time the information and the processing Spinellis did is critical. It may be that Greece is able to find a way to motivate people to begin paying more of the taxes they owe and if and when it does they’ll want good information to go on.

This story feels familiar when I think about many technical control schemes I’ve seen, read about and worked on in the past. Technical solutions have their place and the power of good, smart code and connected systems and robust analysis to provide solid and actionable information can be tremendous. But more often than not, more is needed to address the complex causes of piracy on both the micro and macro scales.

You can read the whole, but short, report by NPR’s Planet Money team here.