THE revisions to estimates are perfectly a fact normal in all statistical seriesas the data stabilizes. No harm done to that, because some are included that were not available at the beginning, when they are published later, and when they are published, they may differ somewhat from the estimated data that was extracted from them to estimate a higher indicator, which adds several like that. other intermediate indicator.
This occurs in national accounting, in which, over time, up to almost four years, since there are revisions up to three years after the first publication, a higher level of disaggregation is available, until completely, forming the tables. input-output of the economy for that year, with the interaction between all branches of activity.
Therefore, reviews are not only commonplace, but are also part of the data collection process. Likewise, it is also normal that, based on these new data, in the revisions, the estimates are slightly corrected, either upwards or downwards. So far, as I said, everything is normal.
Now, what is no longer so normal, but rather exceptional, surprising and extraordinary, is that, from the initial data, the examinations carried out on this or that exercise come to suppose one third variation from the initially published data. If this happens, there is an anomaly, with only two alternatives: either it is intentionally and unorthodoxly modified, or it is very poorly estimated from the start.
The first option should never happen, because statistics must be rigorous to be able to trust the reliability of the estimates and the seriousness of an economy. The second, being less bad, is still something unpresentable, because if the variation is of the order of a third, how was it initially estimated? What guarantee of good estimates would bring a change of such magnitude?
And an equally important modification is what happened with the national accounting estimated by the INE: 2021 data has been changed by more than a third (34%)from the initial 5% to 6.7% during the last extraordinary review of the series. The 2022 review was modified by seven tenths, or a change of 12.73%, which is also excessively large, even if it is the 2021 review that is completely unreasonable.
So, for a few years now, GDP Revisions They are very deep and always in the same direction, with substantial changes, especially in 2021, as I said, with a ripple effect upwards on the whole series. As I said, the fact that the data can undergo a revision is normal and usual, within the limits of statistical rigor, but in small magnitudes.
I have great respect for the INE, which enjoys great prestige and has the clear support of Eurostat, but its independence has been threatened during this six-year term by the attacks received from the government regarding its work. Thus, the Executive, through Calviño, began to say that the INE did not measure either economic activity or the CPI well, at a time when the GDP was showing signs of weakness in its growth and prices were beginning to fall, beyond double digits of the interannual rate. They went so far as to say that they had to review their econometric models, their estimation models. They fought the INE figures with greater assiduity in the publications dependent on the Ministry of Economy itself, in their synthetic indicators. They questioned the work of Juan Manuel Rodríguez Poo, then president of the INE, a technician with a perhaps left-wing ideology, but professional and rigorous, since he was for many years the head of the Institute of Statistics of Cantabria, a professor at the University of Sérida Cantabria, whom I met when I was general director of the Institute of Statistics of the Community of Madrid.
Ultimately, the attacks on the institution provoked Rodriguez Poo has resignedbut having been clearly pushed by the Government. After that, the INE, which had enjoyed great prestige since its embryo in the time of Isabel II, grew throughout the second half of the 20th century and into the 21st century, began to turn in its estimates. The revisions of the GDP were more constant and deeper than ever; It is not that revisions are not normal until a fact is established, as I have already said, but they have never been so deep and so constant. The GDP began to grow significantly, adjusted upwards with each revision, with some curious compositions in its growth, such as the rare increases in some quarters of the Variation of Stocks, which complete the upper heading of the Gross Capital Formation; and the CPI underwent a methodological change promoted by Eurostat, nothing to complain about, but coincidentally, it began to move in the direction indicated by the government.
I want to maintain my confidence in the INE, an institution that I love and whose rigor and independence seem essential to me in a serious country, society and economy, but with the Government’s path in other institutions and with the statements of the Executive regarding the INE’s misestimation, the fact of having forced the dismissal of the previous president of the INE and that precisely, with this change, these estimates have radically evolved in the direction that the Government wanted, it is appropriate to ask, invoking two statistical terms, whether it is a coincidence or causality. I would have liked it to be the former, but with this Government, nothing can be ruled out, which would be extremely sad and I hope that it does not happen.
The INE is a very serious organisation and cannot leave any shadow of doubt about the data it provides. We risk statistical prestige of almost two centuries – since the first antecedent of the INE, the Statistical Commission of the Kingdom, created during the reign of Isabel II, on November 3, 1856 – and the rigor of the figures, and the Government must understand that statistical independence is sacred and that there should be no interference or loss of comparison – as happened, for example, in registered unemployment when the Ministry of Labor has not yet been able to publish how many permanent employees dismissed are in a period of inactivity and does not add to the lists of registered unemployed – . Sacred independence and accuracy of estimates, which cannot be so far from the data of the latest revisions in their first version.
As I am sure that independence is maintained – I would not like to believe otherwise – we must analyse what went wrong for the gap to have been so resounding, to correct it, for the benefit of rigour and seriousness and, therefore, the INE has to ask itself what went wrong, what happened in such a prestigious institution.
