The lack of a standardized framework to report fiscal multipliers limits comparisons across studies, budgetary items, or countries. Within a unified analytical framework (using a panel of 177 countries), we study how key methodological details affect the size and persistence of fiscal multipliers' estimates. Our baseline results are in line with the existing literature with average cumulative medium-term multipliers of -2.1 (-2.5) for taxes on personal income, 0.3 (1.7) for investment and, -0.5 (1.9) for consumption for advanced (emerging market) economies. However, we show that slight changes in the identification of shocks, based on forecast errors or in the definition of the fiscal multiplier, can artificially increase both the size and decrease the precision of estimates. We also emphasize the importance of accounting for the endogenous dynamic responses of fiscal variables to fiscal innovations by showing that multipliers calculated simply as the output response to fiscal shocks, as it is common in the literature, can potentially bias the results.