Income sustainability coefficient
factor.formula
Return Persistence Model (AR(1)):
in:
- :
The earnings indicator of company j in period t. The optional earnings indicators include but are not limited to: earnings per share (EPS), net profit attributable to parent company shareholders, or operating profit. The selection of specific indicators should be considered based on the research purpose and data availability to ensure that consistent earnings indicators are used under the same analytical framework.
- :
The intercept term of the return regression model of company j represents the expected value of the return in period t when the return in period t-1 is zero. This parameter is usually not the focus when measuring the persistence of returns, and more attention is paid to the autoregression coefficient of returns.
- :
The earnings persistence coefficient of company j, i.e., the first-order autoregressive coefficient of the earnings in period t on the earnings in period t-1, measures the autocorrelation of earnings. This coefficient is the core of this factor and is used to evaluate the persistence and predictability of earnings. The closer the estimated value of $\phi_{1,j}$ is to 1, the higher the earnings persistence, i.e., the greater the impact of the current period earnings on the next period earnings.
- :
The regression residual term of company j in period t represents the part of the earnings change that the model cannot explain. It is usually assumed to be a random error term with a mean of 0 and is independent and identically distributed.
factor.explanation
The earnings persistence coefficient $\phi_{1,j}$ is a key indicator. The higher its value, the stronger the persistence of the company's earnings, that is, the higher the autocorrelation of earnings. When $\phi_{1,j}$ is close to 1, it means that the company has a high degree of sustained profitability, and the current period's earnings level can better predict the next period's earnings level, which is generally considered to be a reflection of high earnings quality. On the contrary, when $\phi_{1,j}$ is close to 0, it means that the company's earnings vary greatly, are less persistent, and are difficult to predict. This may indicate that the company's earnings source is unstable, or that the earnings are severely affected by one-off factors, resulting in low-quality earnings levels. This factor can assist in identifying companies with sustainable profitability, help investors make more informed investment decisions, and also help with risk management and identify companies with large earnings fluctuations. In practical applications, panel data regression methods are usually used to regress the same company on a time series, or to perform regression analysis on cross-sectional data to obtain a more robust estimate of the earnings persistence coefficient. In regression analysis, it is necessary to pay attention to potential heteroskedasticity and autocorrelation problems, which can be corrected using robust standard error estimation methods.