Ovaj rad bavi se utjecajem odabranih varijabli na proračunsku transparentnost lokalnih (regionalnih) jedinica. Proračun je jedan od najvažnijih izvora informacija koji omogućuje pregled trenutnog financijskog stanja promatranih jedinica. Kako bi se provjerile zadane pretpostavke provodi se Poissonova regresija. U radu se koriste odrednice proračunske transparentnosti 428 općina, 127 gradova te 20 županija. Odabrane varijable su političari izabrani u prvom krugu izbora, ljevica na vlasti, koalicija na vlasti, visokoobrazovani političari, dob političara te deficit/ suficit. Pretpostavke rada su da varijable političar izabran u prvom krugu i koalicija na vlasti negativno utječu na transparentnost, a varijable ljevica na vlasti, dob političara i deficit pozitivno utječu na transparentnost. Na temelju dobivenih rezultata može se zaključiti da ljevica na vlasti pozitivno utječe na transparentnost općina, dok se ostale varijable nisu pokazale statistički značajnim.This thesis deals with the influence of selected variables on the budget transparency of local (regional) units. The budget is one of the most important information that allows you to review the current financial position of the observed units. In order to check the default assumptions, Poisson’s regression is performed. This paper uses budget transparency constraints for 428 municipalities, 127 cities and 20 counties. The selected variables are politicians elected in the first round of elections, the left-wing government, the coalition of senior, educated politicians, the age of politicians, the deficit / surplus. The assumptions of the thesis are that the variable "politician elected in the first round" and the "ruling coalition" negatively affect transparency, the "ruling" variable, the age of politicians and the deficit positively influence transparency. On the basis of the obtained results, it can be concluded that the leader in the government has a positive influence on the transparency of the municipality, while the other variables have not been shown as statistically significant