Long-term social security statistical forecasts produced and disseminated by the Brazilian
government aim to provide accurate results that would serve as background information
for policy decisions. These forecasts are being used as support for the government’s proposed
pension reform. However, the reliability of official results is uncertain since no systematic
evaluation of these forecasts has ever been published by the Brazilian government
or anyone else. This work aims to present a study of the accuracy and methodology of the
instruments used by the Brazilian government to carry out long-term actuarial forecasts.
More specifically, this work investigates what would be the source of data, assumptions,
equations, variables, parameters and estimation methods used to compute results released
by the federal government. An empirical analysis shows that the long-term Social Security
forecasts are systematically biased in the short term and have significant errors that
render them meaningless in the long run. In addition, attempts to reproduce the results of
the 2012 and 2018 LDOs demonstrate the lack of transparency in official documents, both
in the described equations and the databases used. From a mathematical model developed,
it was shown that the long-term forecasts of variables, such as GDP, Social Security
revenues and expenses, have a large component of volatility and uncertainty which make
your forecast challenging in the short term and impossible, with an acceptable level of
confidence, in the long run. A sensitivity analysis for the productivity and average contribution
rate parameters showed the impact of these on Social Security results, showing a
gain of up to 72% in revenue for an annual labor productivity growth of 3%, for example.
Finally, a free and open source software, developed under this research, that implements
the current official forecast model is presented, as well as several improvements in the
design process, such as the ability to simulate changes in the labor market.
|
