# Census 2022, a sunk cost? (_This text was co-authored with Arilton Ribeiro. It was [originally published](https://www.estadao.com.br/opiniao/espaco-aberto/censo-2022-um-custo-afundado/) in the Brazilian newspaper O Estado de São Paulo on February 19, 2023_) In economics, sunk costs are resources used in the construction of assets that, once realized, cannot be recovered in a meaningful way. This explains that the opportunity cost of these resources, once used, is close to zero. The 2022 Census is at a moment of uncertainty. With six months of data collection, there have been changes in enumerators, neglected outreach strategies, and confusing questionnaires that do not collect as much information as desired. The scenario is worrying, but what matters is the credibility of the collected data, which will be used in the most important decisions of the country in the next ten years. The execution of the 2022 Census has an estimated cost of R$ 2.3 billion, a significant investment. But the costs of an imprecise and unreliable Census are infinitely higher. Census data are fundamental for making strategic decisions about the allocation of resources in the public sector and in private enterprise. Investments in infrastructure, education, health, and social assistance are defined according to the distribution of the population and its exposure to situations of socioeconomic vulnerability. In private enterprise, census data are used to calibrate market research and business strategies. In these cases, the use of unreliable data can generate problems of scale, inadequate business strategies, and the results go far beyond the misallocation of resources. This can even affect the reliability of voting intention polls, with deleterious effects on the quality of public debate and trust in electoral institutions. If the information published by the Census is incomplete or incorrect, there may be an underestimation of needs in certain regions, which can lead to an inadequate distribution of public resources. Imagine that more than 90% of Brazil's municipalities survive through the Municipal Participation Fund (FPM) and that the 2022 Census will determine the revenue bracket for each one. For this reason, there can be no doubt about the quality of the census data, which is why the audit of the data already collected is essential, as former IBGE president Roberto Olinto has highlighted. Conducting an external audit of the results of the 2022 Census, however, is a conjunctural issue. The challenges of carrying out the current Census should raise a deeper public debate on issues of a structural nature, such as the census model adopted and the adoption of institutional improvement measures that allow the country to avoid the problems faced by IBGE in recent years. As for the model, it is necessary to discuss whether the country should not adjust to a trend observed in advanced democracies, abandoning the traditional model based on exhaustive field surveys – adopting a census model based on administrative records or a model that complements information from records with field surveys, by sampling or exhaustively. Currently, the Census collection follows the same format as 50 years ago, with enumerators trained to apply questionnaires that were previously on paper and are now on a technological device. However, Brazil has a huge network of public agents distributed throughout the national territory who recurrently collect data from the population, such as health agents and the teams responsible for the Unified Registry (CadÚnico) for social policies. The Federal Revenue Service receives a large amount of annual information from about 30 million Brazilians. The Superior Electoral Court (TSE) has an extensive and recurrently updated database. The question that remains is: what is the cost of connecting this data through a secure and accessible information exchange environment or platform? Measures such as the mandatory holding of international executive selections conducted by _headhunting_ firms for the choice of the IBGE president and the creation of a fund with budget resources to finance the execution of the Census (avoiding that operating costs are concentrated in one out of every ten budget years) are options that should be debated by the government and civil society. Brazilian society cannot afford to spend the next ten years without knowing ourselves. Decisions about a possible 2024 Census need to be accompanied by a broad debate on the modernization of its format with institutions responsible for its execution. And who knows, maybe we can take advantage of the first Sunday of October to, in addition to voting, answer a complete Census that reflects all Brazilians?