The objective of this thesis is to shed light on the ability of multiple linear regression to estimate valuation multiples. The proposed model is motivated by the theoretical concept that the value two assets with equal sequences of cash flow, risk and timing (Ross, 1976) must be equal as well as empirical evidence in favor of a determinate relationship between valuation multiples and financial measures and ratios. Classically,the estimation of multiples have been done by taking the mean of the observed valuation multiples of similar firms selected by an analyst. This thesis break away from tradition and assume that the fundamentals of a firm have a determinate impact on valuation multiples that can be approximated by regression.To carry out the analyses a dataset consisting of the S&P 1500 Super Composite Index is used for empirical tests of the research design. The initial results of the analysis are mixed but indicated that there are room for improvement in the research design. Therefore the research design is respecified and using the insights obtained from carrying out the initial analysis and tests of robustness of the initial model. The improved model is vastly superior to the original models in terms of forecast precision as well as the coefficient of determination. The results are somewhat inconsistent measured across different sectors. The Informations Technology and Financial sectors prove consistently hard to forecast with the chosen forecast parameters. However, the improved models, excluding those sectors, results in consistently impressive average forecast errors across all the estimated valuation multiples. The forecast precision of the respecified model, excluding the Informations Technology and Financials sectors eclipse the results of earlier studies based on peer groups selected according to industry affiliation and fundamentals (Bhojraj & Lee, 2002 and Plenborg et al, 2017). Thus this thesis provides evidence in favor of using fundamentals to estimate valuation multiples. Furthermore the thesis gives further backing to the notion that industry is a significant predictor of valuation multiples even after controlling for fundamentals as asserted by Plenborg et al (2017). This explains the varying results across sectors. The research design has some advantages over the peer group methodology. First and foremost among these is the relative ease of applying the model. Second is that the method is not limited to work only for firms with a number of close comparables.The disadvantage of the research design is however, that it’s rather poor at describing the sectors Informations Technology and Financials.The analysis is carried out within the constraints of a number of delimitations regarding geography, time and choice of valuation multiples. Thus further studies may back the results beyond these delimitations. The methodology may have several applications in practice although the analysis has emphasized the academical feasibility.
|Educations||MSc in Finance and Accounting, (Graduate Programme) Final Thesis|
|Number of pages||123|