This study compared the characteristic differences between earnings components based on the traditional method and "time-specific" earnings components based on the Francis and Smith (2005) method. Traditional earnings components include not only components that affect current-period earnings but also components that affect prior-period or next-period earnings. Francis and Smith (2005) presented evidence that when components affecting prior-period or next-period earnings are removed, the persistence difference between operating cash flows and accruals is substantially reduced. This study compared the two types of earnings components focusing on their association with earnings, first-order autocorrelation, volatility, and persistence, among the metrics representing earnings characteristics. The main empirical results are as follows. First, according to the cross-sectional distribution of earnings components by earnings interval, the association between accruals and earnings was higher than the association between operating cash flows and earnings among traditional earnings components, but in the Francis and Smith (2005) method, operating cash flows showed a higher association with earnings than accruals. Second, the Francis and Smith (2005) earnings components exhibited higher first-order autocorrelation compared to traditional earnings components. Third, the volatility of the Francis and Smith (2005) earnings components was smaller than that of traditional earnings components. Fourth, the persistence of operating cash flows was smaller in the Francis and Smith (2005) method than in the traditional method, while the persistence of accruals was greater in the Francis and Smith (2005) method than in the traditional method. As a result, similar to Francis and Smith, the difference in persistence between operating cash flows and accruals was substantially reduced. These results demonstrated that the time-specific earnings components based on the Francis and Smith (2005) method possess more stable characteristics compared to traditional earnings components. In other words, this suggests that the relatively lower stability and persistence of traditional earnings components is a phenomenon that arises due to items related to prior-period or next-period earnings that are included in traditional earnings components. Therefore, the results of this study suggest that information users need to take into account these characteristics of earnings components when utilizing them for corporate analysis, including the time-series properties of earnings and earnings quality.