The estimate revisions are a perfectly fact normal in every statistical seriesas the data settles. Nothing to object to this, because some are included that were not available at the first moments, when published later, and when published may differ somewhat from the estimated data that was taken from them to estimate a higher indicator, which adds several such as that other intermediate indicator.
This happens in national accounting, in which, as time goes by, up to almost four years, since there are reviews up to three years after the first publication, a higher level of disaggregation is available, until the same 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 a common occurrence, but 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, upwards or downwards. Up to this point, as I say, everything is normal.
Now, what is no longer so normal, but rather exceptional, surprising and extraordinary, is that, based on the initial data, the reviews carried out on some exercise come to suppose a third of variation on the data initially published. If that happens, there is some anomaly, with only two alternatives: either it is being modified in an intentional and unorthodox manner or it is being estimated very poorly from the start.
The first option should never happen, since statistics must be rigorous 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 magnitude of a third, how was it initially estimated? What guarantee of good estimates would a modification of such caliber provide?
And such an important modification is what has happened with the national accounting estimated by the INE: The 2021 data has been modified by more than a third (34%)from the initial 5% to 6.7% in the last extraordinary review of the series. That of 2022 has been modified by seven tenths, which is a 12.73% modification, also excessively bulky, although it is that of 2021 that is out of all logic.
So, for a couple of years now, GDP revisions They are being very deep and always in the same direction, with substantial modifications, especially in 2021, as I said, with an upward drag effect on the entire series. As I say, the fact that the data may undergo a revision is normal and usual, within statistical rigor, but in small magnitudes.
I have great respect for the INE, which has great prestige and enjoys 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 when prices were beginning to decline. skyrocket, exceeding double digits in the interannual rate. They went so far as to say that they had to review their econometric models, their estimation models. They combated the INE figures with greater assiduity in publications dependent on the Ministry of Economy itself, in their synthetic indicators. They questioned the work of the then president of the INE, Juan Manuel Rodríguez Poo, a technician with possibly left-wing ideology, but professional and rigorous, as he was for many years at the head of the Institute of Statistics of Cantabria, a Professor at the University of Serious Cantabria, whom I met when I was general director of the Institute of Statistics of the Community of Madrid.
In the end, the attacks on the institution caused Rodríguez Poo resignedbut having been clearly pushed by the Government. After that, the INE, with great prestige since its embryo in the times of Isabel II, increased throughout the second half of the 20th century and so far into the 21st century, began to rotate in its estimates. GDP revisions 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 never were so deep and so constant. The GDP has started to grow significantly, adjusted upwards in each review, with some curious compositions in its growth, such as the rare increases in some quarter of the Variation in Inventories, which complete the upper heading of the Gross Capital Formation; and the CPI underwent a methodological change promoted by Eurostat, nothing to object to, but coincidentally it began to move in the direction that the Government had indicated.
I want to maintain my confidence in the INE, an institution that I am fond of and whose rigor and independence seem essential to me in a serious country, society and economy, but with the Government’s track record in other institutions and with the Executive’s statements regarding that the INE estimated poorly, the fact of forcing the dismissal of the previous president of the INE and that precisely, with that change, these estimates radically moved in the direction that the Government wanted, it is worth asking, invoking two statistical terms, if it is a coincidence or causality. I wish it were the first, but with this government nothing can be ruled out, which would be immensely sad, which I hope does not happen.
The INE is a very serious organization and cannot allow there to be even a shadow of doubt about the data it offers. We risk statistical prestige of almost two centuries – since the first antecedent of the INE, the Statistical Commission of the Kingdom, created under 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 neither interference nor loss of comparison – as, for example, has happened in registered unemployment when the Ministry of Labor has not yet been able to publish how many discontinued permanent employees are in a period of inactivity and do not add to the lists of registered unemployment – . Sacred independence and correctness in the 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 want to believe otherwise – we must analyze what went wrong so that the deviation has been so resounding, to correct it, for the benefit of rigor and seriousness and, therefore, the INE has to ask itself what has gone wrong, what has happened in such a prestigious institution.
