Academic journal article Australasian Accounting Business & Finance Journal

Rethinking the Lintnerian Linear Valuation Model

Academic journal article Australasian Accounting Business & Finance Journal

Rethinking the Lintnerian Linear Valuation Model

Article excerpt

(ProQuest: ... denotes formulae omitted.)

1. Introduction

The Ohlson (1995) accounting valuation model has received considerable attention due to its significance in assessing stock price. In his famous model, Ohlson does not set any constraint for the dividend behaviour. Later, Callen and Morel (2000) apply the Lintner (1956) dividend model to the Ohlson (1995) model and develop the Lintnerian linear accounting valuation model (henceforth, the CM model). Lin et al. (2005) extend the Ohlson (1995) model and include dividend information to the valuation model. Their result shows that dividend information has value relevance to a firm?s equity price.

However, Bauer and Bhattacharyya (2007) show that the Lintner (1956) dividend model has several problems. Firstly, according to Bond and Mougoue (1991), if earnings follow an autoregressive process, the Lintner?s dividend process will be an inappropriate process to model a firm?s dividend policy. Since the CM model applies the Lintner model to the Ohlson model under the assumption that earnings follow an autoregressive process, the CM model suffers from the logical paradox. Secondly, if a firm has its goal of payout rate and wants to achieve it, the targeted payout ratio of the firm will be realised eventually. However, it is difficult to ask the Lintner?s dynamic dividend behaviour to stop at the goal of payout rate. Thus, Bauer and Bhattacharyya (2007) establish an alternative dividend process to describe the dynamic behaviour of dividend policy. In this study, we apply the dividend model of Bauer and Bhattacharyya (2007) to the Ohlson (1995) model and develop a new valuation model which is superior to the Ohlson (1995) and CM models empirically.

Prior researchers usually used cross-sectional or time-series approaches to test accounting valuation models. The former method, which focuses on the fundamental values and simultaneously tracks the stock prices and returns, is more popular in the literature (Abarbanell & Bernard 2000; Dechow et al. 1999; Francis et al. 2000; Penman & Sougiannis 1998). However, it encounters a practical limitation due to the time-series nature of the accounting valuation model. Thus, recent empirical studies (Ahmed et al. 2000; Ballester et al. 2002; Callen & Morel 2000) have adopted the time-series approach, examining the time-series relation among equity price, book values, earnings and other value-relevant variables.

However, the time-series approach still has potential drawbacks. For example, Granger and Newbold (1974) and Phillips (1986) have shown that the ordinary least squares (OLS) regressions in nonstationary time-series data could generate spurious results. Therefore, accounting valuation models may also suffer from these problems. Qi et al. (2000) conduct the unit root test of Phillips and Perron (1988) to 95 US firms and find that market values for most of the sample firms are nonstationary. To avoid misleading interpretations of the OLS regression model, it is important to verify the cointegration of accounting variables with equity value. In this paper, we use the Engle and Granger (1987) test to examine cointegration and stationarity of the variables in three different valuation models first, and then compare the effectiveness of the Ohlson, CM and our models.

Karathanassis and Spilioti (2003) employed panel data analysis to compare the explanation ability of the Ohlson model with that of the dividend valuation model. The panel data approach not only renders more efficient and unbiased estimators but also allows more degrees of freedom for estimation. Similarly, we use the panel data approach in this study. According to the result of 1,564 firm-year panel data of US companies from 1973 to 2006, all three models have long-run equilibrium relations. However, our model, which follows the Bauer and Bhattacharyya (2007) dividend model and applies more complicated current earnings, cointegrates for almost 100 percent of sample firms and shows superior ability to describe equity value. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.