23 January 2013

Greek Tragedy

The Financial Times reports today that three statisticians who worked for official Greece statistical agency (the Hellenic Statistical Authority or Elstat) have been charged with crimes for falsely calculating Greek debt in 2009:
Greece has brought criminal charges against the official responsible for measuring the country’s debt, thereby calling into question the validity of its €172bn second bailout by the EU and International Monetary Fund.

Andreas Georgiou, head of the independent statistical agency Elstat, and two senior officials are accused of undermining the country’s “national interests” by inflating the 2009 budget deficit figure used as the benchmark for successive austerity packages.

The three statistical experts face criminal charges of making false statements and corrupt practices, a judicial official said, adding that if found guilty they could serve prison terms of five to 10 years. They have denied any wrongdoing.
The issue appears to involve methodological disputes over how to properly calculate Greek debt, and the choices made by the statisticians in the application of different methods which led to smaller or larger estimates. The calculation of absolute debt levels mattered a great deal politically but also procedurally, as it was a key factor in EU and IMF assistance provided to Greece as a consequence of its financial crisis.

Last April , the New York Times provided some background:
The latest chapter in the complex saga begins in June 2010, a month after Greece signed its first loan agreement with the so-called troika — the European Union, the European Central Bank and the International Monetary Fund — when the former finance minister, George Papaconstantinou, appointed Mr. Georgiou to run the Hellenic Statistical Authority, known as Elstat. 

A year earlier, Greece had been plunged into crisis when the newly elected Socialists announced that the 2009 budget deficit would be 12.4 percent of gross domestic product, twice the previous estimate. In April 2010, the European Union’s statistics agency, Eurostat, revised Greece’s deficit upward again, to 13.6 percent, which forced Greece to seek a bailout. And in November 2010 Eurostat, working with Elstat and Mr. Georgiou, revised the deficit for 2009 upward a final time to 15.4 percent, leading the troika to demand additional budget cuts of $7.65 billion.

How that final calculation was conducted is now the subject of intense debate. Mr. Georgiou has said that it reflects Greece’s first-ever adherence to accepted European procedures. Yet some critics, including some who were on Elstat’s since-disbanded six-person board, said that Mr. Georgiou had actually applied standards that were stiffer than European norms, then tried to thwart them when they raised questions about the process.
Looking a bit deeper the NYT explained that the methodological process for calculating debt levels in Greece and the EU more broadly was less an actuarial exercise than a politically-negotiated social construction:
Mr. Georgiou said that members of the board had the incorrect assumption that they could “vote” on methodology and statistics, and that some represented vested interests that did not want Greece’s dire finances to come to light. “We were faced with significant pressures through the board not to revise the deficit upwards on account of fully applying European Union rules, but to minimize it,” he said.

In the past, countries in a stronger position than Greece have traditionally negotiated with Eurostat over how to classify items in the government debt. In testimony before the parliamentary committee, other Greek officials said the country had lost that ability once it accepted foreign aid.
Walter Radermacher, the director of Eurostat, the EU statistical agency, provided this bit of wisdom:
The truth is not my business. I am a statistician. I don’t like words like ‘correct’ and ‘truth.’ Statistics is about measuring against convention.
Georgiou offered his explanation of goings on in 2011:
I am being prosecuted for not cooking the books. We would like to be a good, boring institution doing its job. Unfortunately, in Greece statistics is a combat sport.
This case will be resolved under Greek law for which I can offer no expertise. However, this case shares some interesting similarities and differences with the recent judgement against Italian government scientists for their role in providing what was determined to be misleading and inaccurate information in advance of the L'Aquila earthquake.

This week the judge in the L'Aquila case provided his justification for convicting the scientists, explaining that they had misled the public and did so in collusion with politicians, The judge explained that, in effect, the scientists were guilty of delivering a politically-friendly message that was not in fact well-supported by the state of the science, contributed to poor decisions by L'Aquila residents.

In contrast, the Greek statisticians have been charged for what appears to be the opposite crime -- they failed to bend to the will of politicians and instead delivered a politically-unfriendly message. The statisticians did so in the context of statistical conventions that are not well-established or agreed upon.

A third recent case that is relevant is the so-called "hurricane deductible" associated with insurance payouts related to "Hurricane" Sandy. In that case, immediately after the storm politicians quickly moved to define conventions for insurance payouts in a manner that best suited their desired political outcomes, regardless of what the science may say about Sandy's meteorological status at landfall. In the process, the politicians defeated any chance to share costs of insurance according to risk of property location using a scientific metric. The final scientific judgments are not in on Sandy, and politicians have made clear how they think such judgments should be made by the responsible government agency.

What ties the three cases together is a lack of strong institutions able to arbitrate empirical questions independently. The lack of such institutions means that science -- whether hurricane, earthquake or economic -- played little role in these decisions. Yet in each case there was the expectation that input from independent experts might contribute to improved decision making. In each case arguably the opposite occurred.

The failures here have nothing to do with anyone being "anti-science" but rather stem from poor policy design where expertise meets politics. In the case of the Italian and Greek experts the outcomes are personally tragic. More broadly, in each case the public suffers from the loss of expert knowledge in decision making. Such situations are entirely preventable.