Valuation por múltiplos: seleção do múltiplo, comparáveis e cálculo

Abstract

The use of multiples for the valuation of companies is a broadly practiced methodology in the market, individually or complementing a valuation performed through the discounted cash flow method (Fazzini, 2016; Liu, Nissim and Thomas, 2007; Schueler, 2020). A study conducted by Bagna and Ramusino (2017) found, through the Mean Absolute Percentage Error (MAPE) presented, the precision of multiples from cyclical european companies calculated through the arithmetic mean, median and harmonic mean, considering peer selection through industry membership. Additionally, R. G. Serra and L. P. L. Fávero (2018) indicated that the grouping of comparables by similarity of economic fundamentals would be preferrable to the grouping by industry membership in a context of a homogeneity analysis of multiples in a group. Based on the studies previously mentiones, the objective of this work was to verify, based on MAPE, the precision of multiples EV/EBITDA, P/B and P/E calculated for brazilia companies listed on B3 through the arithmetic mean, median and harmonic mean, also verifying if there is a relevant variation in the error found considering the grouping of comparables through economic fundamentals in comparison to grouping by industry membership. For this, companies listed on B3 were analyzed, in the period from 2011 to 2020, considering only companies which presented all necessary data and positive multiples. With this data, multiples EV/EBITDA, P/B and P/E were calculated, through arithmetic mean, median and harmonic mean, considering grouping by industry membership, based on the NAICS sector, and based on economic fundamentals. With this, the MAPE presented by the different calculations of each multiple was verified. The result of the analysis indicated that, for all scenarios, the calculation of the multiple through harmonic mean presented the smaller error. Additionally, in general, Equity multiples (P/E and P/B) presented a superior performance to EV/EBITDA. When considering industry membership peer selection, the P/E multiple presented the smaller error for most of the analyzed period. With the peer selection through similarity of economic fundamentals, considering Equity metrics, the P/B multiple presented the best performance in 50% of the time. When selecting peers through similarity of economic fundamentals considering firm metrics, the P/B multiple’s superior performance becomes dominant, occurring in 9 out of 10 of the analyzed Years. This work contributed by indicating, to professionals in this field, what multiple, mean and comparable grouping produces the smaller estimation error