Paul Alper points us to this news article by Catherine Rampell, which tells this story:
Georgiou is not a mobster. He’s not a hit man or a spy. He’s a statistician. And the sin at the heart of his supposed crimes was publishing correct budget numbers.
The government has brought a relentless series of criminal prosecutions against him. His countrymen have sought their own vengeance by hacking his emails, dragging him into court, even threatening his life. His lawyers in Greece are now preparing for his latest trial, which begins this month . . .
Politicians accused him of being a “Trojan horse” for international interests that wanted to place Greece under “foreign occupation.” It didn’t matter that his numbers were repeatedly validated by outside experts. Or that the deficit his agency calculated precisely matched the net amount Greece borrowed from capital markets in 2009.
The government prosecuted, cleared and re-prosecuted him anyway, for causing “extraordinary damage” to the Greek state and for “violation of duty.” In one case, he was given a suspended prison sentence of two years. Two criminal investigations remain open.
I’m reminded of this story, The Commissar for Traffic Presents the Latest Five-Year Plan. There sometimes seem to be incentives to give inaccurate forecasts that tell people what they want to hear.
Getting back to the Greek story, Alper writes:
Consider yourself personally lucky—Wansink, Brooks, Bem, etc.—that you don’t live in Greece because:
In layman’s terms, a court said he made statements that were true but that hurt someone’s reputation. (Yes, this is an actual crime in Greece.) If his appeal fails, he’ll be forced to pay and publicly apologize to his predecessor. This means the person who restored the credibility of Greek statistics will have to apologize to a person who had been fudging the data.
Wow. I guess whistleblowers have it hard there too.
Alas, everything is paywalled, nothing to help with my pandemic fatigue.
NB: This is a comment about the “Commissar for Traffic” story, rather than the harrowing Greek statistician story.
The end of the “Commissar” article describes a tension between getting the right answer and using conventional methods, and I think this is important. I also think it is an instance of the more general issue that a good model needs to balance predictive accuracy with communicability.
Normally, we would say it’s a balance between “fit” and “complexity”, and this would be quantified using an information criterion or Bayes factor or cross-validation metric. But I think “communicability” better captures the dimensions of complexity that are relevant in both applied and theoretical modeling. In a sense, this is a form of algorithmic complexity, but instead of being defined in a formal language, it is in the natural language of a domain.
A model that adheres to convention is easier to communicate. Why can this be a good thing? Because it helps someone understand *why* the model makes the predictions it does, it reveals the causal mechanisms that the model is supposed to represent. This also makes it easier to figure out why it makes wrong predictions, which might even mitigate against a desire for perfect “fit”.
Of course, consumers of modeling need to be able to learn too, such that when conventional methods really can’t deliver, they can add new words to their language. Another reason why I like the term “communicability”: communication is a two-way street (to bring it back to the traffic example).
By “communicability” do you mean something different than inherent interpretability?
(Familiarity can often be a form of miscommunication.)
https://gelmanstatdev.wpengine.com/2019/11/15/zombie-semantics-created-in-the-hope-of-keeping-most-on-the-same-low-road-you-are-comfortable-with-now-delaying-the-hardship-of-learning-better-methodology/
What I’m talking about is essentially the idea of algorithmic complexity, but with reference to the actual language used to convey a model rather than an abstract formal language.
Just like the amount of information conveyed depends on whether the sender and receiver have the same codebook, the “complexity” of a model depends on the language. If the modeler and the consumer/reader have different background knowledge, the won’t have the same terms in common. As a result, the model will be more “complex” because it will take longer to explain. That’s why I like the term “communicability”.
It also emphasizes that “inherent interpretability” is “inherent” to the person, not the model. As you say in your post, that means that there is an onus on people (both modelers and consumers of modeling) to learn new terms/techniques when the existing ones are inefficient or misleading. But it also helps us understand why people might stick with zombie terms past the point of their usefulness–because they think using them aids communication.
Very good points.
>Politicians accused him of being a “Trojan horse”
I’ll defer to the Greeks on this one, they should know
“I’ll defer to the Greeks on this one”
Wouldn’t the Greeks call it a “Greek Horse?”
If using a Trojan horse is seen as a bad thing, they’d be happy to blame the Trojans! Especially since they are extinct.
Well, the original Trojan horse was used by the Greeks to sneak into Troja.
That is the perfect handle for a response to this post.
Article 362 has to be read together with Article 366 of the Greek Criminal Code. If he is reporting fact, as CLEARLY stated in 366, then the acts of 362 cannot be punished.
This is terrible, to just read 362 without a reference to the “General Provisions” of the chapter.
Mp:
It could well be that when all is said and done the courts will decide that there was no case against Georgiou. In the meantime, though, it’s horrible that he’s been repeatedly prosecuted for the crime of telling the truth.
“There sometimes seem to be incentives to give inaccurate forecasts that tell people what they want to hear.”
Anyone remember Sharpiegate?
The statistics & economic community should support this fellow. (I am not a member of either community.)
Significance of August 2017 and May 2018 ran coverages on Georgiou’s woes. As did The Economist of 06 June 2018, as well as the American Statistical Association in a November 2019 statement. There were similar concerns voiced by EU committees like the European Statistical Advisory Committee on September 2017 and on March 2019.
This is indeed horrible. Is any professional organization putting together a petition drive about this?
Several professional associations and international organisations have protested against what has been rightly called the persecution of Georgiou. Here is a link to a recent statement by ASA. At the end of the statement there are various links. The most detailed one describing the background and history of the affair is : http://uti.is/2020/04/the-georgiou-affair-how-greece-keeps-failing-the-political-corruption-test/
The link to ASA’s statement is: https://www.amstat.org/ASA/News/ASA-Reissues-Georgiou-Statement-on-Mitsotakis-One-Year-Anniversary.aspx
Yes, and also see the earlier statement signed by many statisticians and organizations: https://www.amstat.org/asa/News/ASA-Statement-Prominent-Individuals-Urge-Greek-Authorities-to-Halt-Prosecutions-of-Andreas-Georgiou.aspx