本研究採用美國季資料探討Ohlson(1995)評價模型是否成立,在過去的統計分析程序下,這些模型中所有變數都假設為定態,而一般實證研究也皆假設為定態,然而這些變數可能並未符合此假設,因此迴歸分析可能會導致虛偽迴歸的結果,故使用單根檢定和共整合檢定來測試Ohlson模型時間序列特性。傳統的共整合檢定不能解決power太低的問題,因此許多學者採用追蹤資料改進power的問題且Panel單根檢定和Panel共整合檢定可應用於更長時間序列和更大的橫斷面資料。 在本研究中採用追蹤資料驗證 (1)市場價值、帳面價值和異常盈餘是否為定態。(2)帳面價值和異常盈餘與市場價值是否存在長期的均衡關係。
This study uses American quarterly data to examine whether Ohlson(1995) model is holding. In the past, the statistic analyzing processes assume all variables stationary for the regression. Generally empirical studies assume that variables are stationary, but most variables do not conform to the assumption. Therefore, non-stationary variables may result in spurious regression. We will use the unit root and cointegration test to examine time-series properties of the Ohlson model. Conventional cointegration method can’t solve the problem of power being too low. Some studies have adopted panel data in order to improve the power. However, panel-unit root test and panel-cointegration test are applied with long time-series and large cross-section data. In this study, we use panel data to test that: (1) whether market value, book value, and residual income are stationary and (2) if they are nonstationary, whether book value and residual income cointegrate with market value.