Caesar's wife should be alive and well, and living at the NSO
Unless politicians, as well as non-partisan analysts, accept that there will be revisions, and treat statistics with a degree of caution, there will be endless suspicion and controversy.
Parts of the media are reporting that, according to the Prime Minister, the economy (as measured by the Gross Domestic Product) grew by three per cent in real terms during 2006. Presumably Dr Gonzi is fully aware that, by making such statements, he gratuitously feeds the controversy that has erupted over the way the National Statistics Office (NSO) works. Aside from the way the outgoing director of the NSO was turned into a political and administrative football, leading to his resignation, the controversy turned technical issues on their head.
The statistics office had to revise data, including those relating to GDP, going back 10 years. That was necessary in order to bring the methodology of the data in line with that used by Eurostat, the statistics authority of the European Union. The recasting and updating process had begun in 2003, over two years before the helm of the NSO passed into a new pair of hands.
It will be completed next month. The revision in the NSO time-series affected by the compliance requirement will not change the reality experienced over the 10 years involved. If it is true that what will be, will be, it is truer still that what has taken place cannot be changed or any part of it erased. The controversy of the revision was totally unnecessary. Controversy over the correction by the NSO of the error that corrupted the statistics to boost the deficit for 1996 well beyond the figure that had been recorded for years, is more understandable.
That is not to say it was fully justified. To err is human. But it is also human for politicians to be suspicious. When errors come to light it is important that explanations are given why they occurred, and to ensure that they not recur. The NSO should be above controversy. There will be enough controversy in the way the official statistics are interpreted. Nor will it ever be the case that everyone will understand or agree with what the statistics tell.
If the statistics record an increase in real GDP, there may still be sectors which suffered a decline. Also, very often the official statistics refer to averages. The lower one stands than the average, the more difficult will it to be to understand what exactly are statistics all about.
Leaving perception aside, unnecessary controversy that casts doubt over the veracity of official figures hinders proper appreciation of what is taking place, and thereby constraints policy decisions which need to be made on an ongoing basis on the data in hand, and forecast that start from them.
At this stage it is important to reduce the heat and the smoke that are engulfing the National Statistics Office. It is enough that the NSO has lost the services of a very able person because of them. The country cannot afford to lose the effective services of the NSO itself, which is what will happen if its output is not accepted as being valid and as above suspicion as Caesar's wife.
It will not be easy for that to happen, while the political parties cynically fire away charge and counter charge. There is a technical factor which makes it still more difficult. Several of the data compiled by the NSO will be kept "open" for three years. When they are published, they will be "provisional" and not final. There are data that become final when they are released, or soon after, like those on consumer and retail prices. But other aggregates, including exports, imports and the GDP, which they influence directly, will be kept "open" for three years.
That does not mean they will be wildly off the mark. The NSO's methodologies, like those of the Central Bank, are updated and improved on an ongoing basis. But, over the period before they are "closed" and made final thereby, there will be revisions. Unless politicians, as well as non-partisan analysts, accept that there will be revisions, and treat statistics with a degree of caution, there will be endless suspicion and controversy.
The political parties' claim to have technical expertise to assist them, even if they may not always be able to reveal their identity, should make it possible for them to engage in healthy critical exchange of views. Thereby, they can contribute towards ongoing technical upgrading for the good of present and future decision makers, instead of feeding sterile controversy.
By rushing to state that GDP went up by three per cent last year, Prime Minister Gonzi merely invited more controversy. The NSO has not spoken yet. Why did the PM jump its gun? Moreover, when the NSO does publish the GDP data for 2006 - adding its estimate for the fourth quarter to those already published for the first three - it will surely stress that the data are provisional. It would also be wise if, henceforth, it starts pointing out the point from which the data is still "open" in the relevant time series.
Politicians who quote provisional and open data without pointing that out, do a disservice to the NSO, rather than helping it to be like Caesar's wife.
The statistics office had to revise data, including those relating to GDP, going back 10 years. That was necessary in order to bring the methodology of the data in line with that used by Eurostat, the statistics authority of the European Union. The recasting and updating process had begun in 2003, over two years before the helm of the NSO passed into a new pair of hands.
It will be completed next month. The revision in the NSO time-series affected by the compliance requirement will not change the reality experienced over the 10 years involved. If it is true that what will be, will be, it is truer still that what has taken place cannot be changed or any part of it erased. The controversy of the revision was totally unnecessary. Controversy over the correction by the NSO of the error that corrupted the statistics to boost the deficit for 1996 well beyond the figure that had been recorded for years, is more understandable.
That is not to say it was fully justified. To err is human. But it is also human for politicians to be suspicious. When errors come to light it is important that explanations are given why they occurred, and to ensure that they not recur. The NSO should be above controversy. There will be enough controversy in the way the official statistics are interpreted. Nor will it ever be the case that everyone will understand or agree with what the statistics tell.
If the statistics record an increase in real GDP, there may still be sectors which suffered a decline. Also, very often the official statistics refer to averages. The lower one stands than the average, the more difficult will it to be to understand what exactly are statistics all about.
Leaving perception aside, unnecessary controversy that casts doubt over the veracity of official figures hinders proper appreciation of what is taking place, and thereby constraints policy decisions which need to be made on an ongoing basis on the data in hand, and forecast that start from them.
At this stage it is important to reduce the heat and the smoke that are engulfing the National Statistics Office. It is enough that the NSO has lost the services of a very able person because of them. The country cannot afford to lose the effective services of the NSO itself, which is what will happen if its output is not accepted as being valid and as above suspicion as Caesar's wife.
It will not be easy for that to happen, while the political parties cynically fire away charge and counter charge. There is a technical factor which makes it still more difficult. Several of the data compiled by the NSO will be kept "open" for three years. When they are published, they will be "provisional" and not final. There are data that become final when they are released, or soon after, like those on consumer and retail prices. But other aggregates, including exports, imports and the GDP, which they influence directly, will be kept "open" for three years.
That does not mean they will be wildly off the mark. The NSO's methodologies, like those of the Central Bank, are updated and improved on an ongoing basis. But, over the period before they are "closed" and made final thereby, there will be revisions. Unless politicians, as well as non-partisan analysts, accept that there will be revisions, and treat statistics with a degree of caution, there will be endless suspicion and controversy.
The political parties' claim to have technical expertise to assist them, even if they may not always be able to reveal their identity, should make it possible for them to engage in healthy critical exchange of views. Thereby, they can contribute towards ongoing technical upgrading for the good of present and future decision makers, instead of feeding sterile controversy.
By rushing to state that GDP went up by three per cent last year, Prime Minister Gonzi merely invited more controversy. The NSO has not spoken yet. Why did the PM jump its gun? Moreover, when the NSO does publish the GDP data for 2006 - adding its estimate for the fourth quarter to those already published for the first three - it will surely stress that the data are provisional. It would also be wise if, henceforth, it starts pointing out the point from which the data is still "open" in the relevant time series.
Politicians who quote provisional and open data without pointing that out, do a disservice to the NSO, rather than helping it to be like Caesar's wife.