Can a Budget Simulation Model of Decentralized Territorial Authorities be Enhanced by an Artificial Intelligence Method

Abstract

Local authorities in Madagascar do not have yet a budget simulation model. Budgetary rationalization makes it possible to effectively predict and manage budget implementation. However, the local finance budget is always based on an ex-ante situation without taking into account the economic, political, the behaviour of the elected official, the electoral cycle, and performance indicators. Therefore, a numerical simulation (use in many developed countries) from computer tools or the use of artificial intelligence or AI that can effectively predict a local finance budget. Predictive analyses help to better identify the fundamental characteristics of the budget elements, in order to model them and anticipate the executions as best as possible. Rather, it is a practice that relies on statistical tools, but also on the search for correlation and regressions that are methods used to discover functional models, connecting a variable to be explained and a set of explanatory variables. The prediction of a primitive budget then consists of using the different models from a simulation in Prex-B-developed software for the budget of local finances or decentralized territorial collectivities.