The Koyck Model in Predicting Public Revenues of the Most Populous Districts in the State of Bahia, Brazil

Abstract

Listen

Translate article

The forecast of public revenue is a key part in the budget process, becoming an instrument that supports public managers for making decisions regarding an appropriate budget execution. This research aims to determine the effectiveness of the Koyck model in predicting public revenues of the most populous counties in the state of Bahia, comparing it to the model adopted by the Federal Budget Secretariat (SOF). For data collection the predictions made by the municipalities surveyed during the year 2000- 2011 were analyzed. This required the implementation of a quantitative research that uses statistics facing econometrics, with documentary feature, which documents the estimates of the budgetary revenues of the municipalities surveyed, analyzed through the site of the Court of Accounts of the municipalities of Bahia (TCM-BA). The survey results showed that the Koyck model provided effectiveness on forecasts of budgetary revenues, compared to the model adopted by municipalities that are more susceptible to errors. The average error percentage of the municipalities for this research, using the methodology proposed by SOF, obtained an average percentage error of 13.21%, compared to the Koyck model that had a percentage of average error of 8.87%. It is observed that the methodology used by Koyck obtained efficacy in 104 (one hundred and four) of 144 (one hundred and forty-four) timeframes. It is concluded that the Koyck methodology is effective in the forecast of public revenue in the most populous counties in the state of Bahia. Therefore the development of new research in the area to improve the decision making of managers and an appropriate estimate of government revenue is necessary.