Use and abuse of official statistics in Latin America: The case of Argentina

Abstract

In line with the Fundamental Principals of Official Statistics to produce valid and reliable statistics Governments need to provide the legal framework and resources to the statistical system to allow statisticians to produce the required statistics, without interference, using the best available methodology and techniques from the most suited sources of information. In Latin America and the Caribbean the colonial past affected and still affects the production of statistics. During the Colonial period statistics were of limited scope and use, mostly serving the interests of the Colonial powers. After independence in Latin America statistics became an instrument for development only after World War II, while in the Caribbean the newly independent nations had to adjust the Colonial system to national souvereignity. Conflicts between statistical independence and administrative desire and convenience did occur. Occasionally statisticians were under pressure to modify results to serve administrative or political purposes. An extreme case of Government interference with statistical activities is the case of Argentine since 2007. The gross manipulation of the Consumer Price Index (CPI) that began at that time was aimed at concealing the rise in inflation which took place at the beginning of that year. Statisticians in the National Statistical Office who refused to be part of the forgery were demoted, and dismissed while others resigned. The alteration of the CPI severely affected other statistical indices. Private consultants and researchers were subject to criminal prosecutions and punished with hefty fines for the “crime” of publishing their own price estimates. Although in most cases the judicial system acquitted them, this happened some years later, and currently there are still researchers awaiting the final judgement. In spite of the reaction by public opinion and the world statistical community nothing changed substantially until now. The paper concludes with some recommendations to safeguard the integrity of statistics inspired by this sad experience.