Towards a more accurate equity valuation

The accurate valuation of companies is of fundamental importance for investors, analysts, underwriters, managers and many others. Capital market participants, for example, make an enormous effort to value firms and identify undervalued stocks. Managers are constantly relying on accurate valuations in order to create shareholder value. For example in M&A transactions, their offer price needs to be close below the intrinsic value of the target in order to outbid competitors and still create shareholder value. At the same time, they face the danger of paying too much. Accounting standard setters such as the Financial Accounting Standards Board (FASB) or the International Accounting Standards Board (IASB) try to support accurate valuations by improving the ability of the financial reporting system to provide information about the amount, timing and uncertainty of future payoffs. Finally, researchers seek to understand and improve the process of price discovery at the capital market. There is a large literature developing, testing and comparing valuation models and models of risk and return. In addition, there is a large literature which investigates the properties of forecasted future payoffs and how these may be improved. Clearly, society spends a lot of effort to achieve accurate valuations of firms.

There are several distinct valuation methods that can be employed to value a company. To structure the different valuation models investigated in this thesis I use Figure 2 and differentiate between models that require explicit forecasts of future payoffs and those that do not require explicit forecasts of future payoffs.

The accurate valuation of companies is a primary concern for anyone involved in valuing firms. However, studies implementing different valuation methods frequently find deviations of more than 50% between value estimates and observed market values. One interpretation of these results is that capital markets are grossly inefficient. A more plausible alternative explanation is that it is difficult to obtain perfect estimates of future payoffs and/or cost of capital. As outlined in Figure 1 in Section 1.2, I therefore try to analyze why valuations are apparently so inaccurate and how they may become more accurate. The analysis in this chapter is based on a review of the empirical literature on the valuation errors of equity valuation models. Such a review of the literature is useful for several reasons. Most importantly, for many practitioners their choice of valuation method depends - amongst other factors - on the assumed valuation accuracy of this method. However, I am not aware of any comprehensive review on the empirical performance of different types of valuation models. Hence, the review presented in this chapter may help researchers and practitioners to assess the relative and absolute performance they may expect from different valuation methods under different settings. With respect to the second research question, I want to investigate which factors affect valuation errors and whether some valuation methods are more accurate than others. Most importantly, I believe that valuation accuracy is not fixed or stationary but rather depends on a number of factors. Therefore, the second aim of this review is to help researchers and practitioners to evaluate which factors affect valuation accuracy and to what extent they do so. Finally, to address the third research question, I derive best practice approaches for different valuation methods.

Multiples are the most common valuation technique. For example, Asquith/Mikhail/ Au (2005) report that 99% of top analysts use a multiplier model for firm valuation and Roosenboom (2007) finds that underwriters typically rely on multiples when valuing initial public offerings. At the same time, multiples are prone to incorrect implementation or even manipulation. The review in Chapter 3 indicates that value estimates may vary significantly when employing a different value driver (e.g. book value of equity instead of earnings) or a different set of comparable firms (the ‘peer group’). It appears that some methods to choose comparable firms perform better than others. Such deviations between value estimates frequently arise, because it is difficult to find a peer group which corresponds to a target firm in all value relevant characteristics. Hence, I try to investigate the factors affecting valuation errors. Overall, this chapter investigates: 1) why and to what extent different value dnvers are affected by differences between firms, 2) how biased or inaccurate valuations may be detected and 3) how valuations based on multiples may be improved to account for differences between firms.

As outlined in Section 2.4.1, intrinsic valuation methods frequently employ a linear information model to predict the stream of expected future payoffs beyond the forecast horizon. Hence, improvements in the implementation of linear information models might help to improve valuation errors for the residual income model in general. In addition, such an analysis is useful to understand how published accounting information relates to future payoffs and observable market values.

This thesis investigates the valuation accuracy of equity valuation methods. The mam aims of this thesis are 1) to empirically analyze the absolute and relative valuation errors of different equity valuation methods, 2) to empirically analyze the determinants of valuation errors and 3) to improve the valuation accuracy of equity valuation methods.