How can you eliminate endogeneity of earnings and dividends to determine if dividends signal future earnings?

Endogeneity is one of the main problems of any economist wishing to study the relationship between several variables. Mathematically, this problem corresponds to the fact that the (central) hypothesis of non-correlation between the explanatory variables (x) and the error term (e) will be violated. This results in a bias in the estimation of the coefficient of interest and therefore fallacious conclusions. There are 3 main sources of endogeneity in econometrics: Simultaneity: if the economist wants to show that x causes y, it is necessary at the same time that the variable x is not influenced by y. See more details on this previous lighting here. In this context one of the solutions is to instrument the variable x by another variable w not influenced by y. Unobserved heterogeneity: the relationship between x and y may simply be related to the effect of a third factor z. See more details on this previous lighting here. In this framework it will be necessary, if possible, to add a variable in the controlling model for this factor z. - Measurement error: some variables may not be measured accurately, which will affect the estimate results in a bias in the estimation of the coefficient of interest and therefore fallacious conclusions [1].

Thanks for invitation,

Agree with our colleague Mr. Fateh Bouchene's answer which I do believe covered the subject.

Here it the importance of the complete audit process to be adhered and follwed to redue the negative or bad effects of the wrongly applied earnig management presentation. 

That is why the general asseamblies fora business firm, should take the great step of responsibilty to hire the approbriate the needd qualified Auidtor