Dr Pali Lehohla
Dr Pali Lehohla

OPINION : The perils and pitfalls of a modern-day statistician

By Dr Pali Lehohla Time of article published Jul 25, 2017

Share this article:

GALILEO Galilei (1564-1642), is an Italian who immersed himself in observation to advance and transform knowledge from postulates to theorems.

He was a scientist of immense rigour who, albeit a Christian, would not be intimidated by priestly threats that too often stood in the path of scientific enquiry.

House arrest and imprisonment would not stop him from articulating a scientific observation that the earth revolves around the sun and rotates on its axis. “E pur si mouve” which translates into “Yet it moves”.

This was a papal punishment against what was seen to be heresy. It was only towards the closure of the 20th century that a public papal apology to punishing Galileo’s “e pur si mouve” was made.

In the previous column I raised the emergent perils confronting a modern statistician. These dangers are not accidental, but emerge as statistics comes of age in systems of accountability.

This is especially as democracy takes hold and is supported largely by the dimension of transparency, which is brought about by an explosion of information technology and the power of own hand-held devices.

Systems of accountability were first articulated by the judiciary. This is well captured by Andreas Georgiou, a former Greek statistician whose unending prosecution by successive Greek authorities since 2010 is a matter of public knowledge.

And it is a matter of peculiar curiosity demonstrating the Galileian perils that confront a chief statistician in modern systems of government four centuries later. He makes reference to Montesquieu’s statement in 1748 on the independence of the judiciary and argues that this principle must apply to official statistics.

In his paper delivered at the 61st World Congress of Statistics titled The production of official statistics needs to be a separate branch of government he argues that “the independence of official statistics has to be real and not apparent merely” consistent with Montequieu's statement.

My informed view on matters of official statistics is that systems of human endeavour dealing with accountability have evolved.

With each milieu they advance an arm of the state into the realm of Montesquieu’s independents. The first to mature and get early recognition was the judiciary through indeed the Montesquieu principle of independence.

Following closely on the judiciary was the legislature as an independent arm of the state, largely driven by the advent of parliamentary democracies that started holding their own in many a jurisdiction.

More recently in the 21st century, supreme audits claimed their independent space as a crucial part of systems of accountability that have to do in the first instance with financial prudence.

So unsurprisingly so, as the second decade of the 21st century comes to a close, largely driven by the advent of information technology and ubiquitous hand-held devices as well as the Sustainable Development Goals, official statistics has increased its visibility and its intrinsic value of a conduit of trust. In this regard they have been catapulted to the high table and thus knocking in this famous space of Montesquieu independents.


Andreas Georgiou is a former president of the Hellenic Statistical Authority, serving from 2010-2015. He is an MIT graduate. He was recruited from the IMF to be the president and at the end of his tenure, he became a visiting lecturer at Amherst College in the US and a consultant in official statistics.

His first task that put him in trouble was to revise the Greek deficit of 2009. The level of deficit was revised to 15.4 percent from 13.6percent in 2010.

These findings, which were undertaken applying Eurostat practices and confirmed by the IMF, revealed that Greece had historically underestimated its deficit.

The government was not pleased, and Georgiou has since faced charges of breach of faith by his government, with members of his board serving as witnesses in this case where he and two others are accused of artificially inflating the deficit.

If convicted Georgiou faces up to five years imprisonment. To his surprise and ours as practitioners of statistics, Georgiou’s findings followed the EU regulations and standards of compilation.

“The statistics we have been accused of ‘çooking’ are the ones that did not receive any reservations from Eurostat”, says Georgiou.

On December 6, 2016, Georgiou was found innocent of all charges proffered by a First Instance Court, but the unanimous acquittal by a three-judge panel was annulled by another prosecutor and Georgiou faced a double jeopardy trial last week on July 18 and the hearing will resume on July 31.

Georgiou has been there before, as in September 2015, he was first acquitted by the Appeals Court Council of any wrongdoing, but the prosecutor in the Supreme Court decided to annul the acquittal on August 1, 2016.

Over the past six years Georgiou, whose daughter was seven when the court case started, has been facing trial on matters of official statistics, making official statistics a high risk career.

When I served as a vice-president of the International Statistics Institute from 2009 to 2013, our board wrote to the Greek parliament in 2013 protesting the persecution of Georgiou.

In March 2012 at the UN Statistics Commission Seminar on “Violations of the UN Fundamental Principles”, I was part of the panel. In my presentation on the list of offenders by omission or commission were Canada - 2010 Census long sample design non-random; Argentina falsification of the Consumer Price Index; Greece falsification of deficit numbers; Indonesia - household privacy violations and South Africa - inflating the Consumer Price Index (CPI) over fifteen months from 2002.

By the way, I was already then the Statistician-General and the inflated numbers were under my watch. What were the countries' reactions?

The Argentinian reaction was most virulent, as they called me a liar and walked in protest out of the presentation where there were in excess of 300 participants.

Their ambassador accosted the South African Permanent Mission and they petitioned the director of the UN Statistics division as well. South African ambassador Baso Sanqu summoned me on the same evening for an explanation.

I also reported this instance to Minister Trevor Manuel as soon as I arrived back home. My response to all these authorities was understood to be impeccably professional, fair and firm.

The Canadian statistician responded as well and argued that the census still remained scientific and my retort was that a volunteer sample cannot be scientific.


What happened in these jurisdictions since these violations?

In South Africa I admitted to error and no amount of apologies could ever be enough given the impact of the error, but I indeed remain gracious to the government and peoples of South Africa for their magnanimity and understanding that errors do and can occur.

I led a gruelling re-engineering of the economic, social and population statistics and the modernisation programme of our statistics system and corrected the 2002 error in the CPI.

That has remained a big blot and a constant reminder in my career as a statistician and as a statistician-general.

In Canada, the 2016 Census has reintroduced a mandatory long form and there is major legislative reform that was occasioned by government action of annulling or tempering with the sample design implementation of the long census form.

The Argentinians have remained very appreciative of my unrelenting stance of critiquing them during their errant days. They are now reforming their statistics and I have been invited to view their work and progress they have made.

The former Greek statistician Georgiou remains my intellectual comrade as we weave through the perils of official statistics and trust that Greece will see the light of day. Indonesia remains stuck with benevolent violation of privacy and the risks remain real.

Like Galileo, the scientist who faced the perils of religion, chief statisticians face the perils of politics. Will the Montesquieu principle of independence save them? The South African Statistics Act 6 of 1999 is now under legislative review.

Dr Pali Lehohla is South Africa Statistician-General and Head of Statistics South Africa.

Share this article: