Being a physicist/engineer and amateur long-term dividend investor (contrarian by my nature) I try to understand mean-reverting in stocks because of importance of this concept . So far I did not figure out fundamental forces for mean-reverting in stock market except only one.
In my opinion P=FV+MP, where P is stock price, FV is the fundamental value and MP is the mispricing. It should be noted that P is always positive and the precise number, while FV and MP can be positive and negative and BOTH are ill-defined numbers, so noise/signal approach known in hardware engineering might not work in finance. Assuming that FV is time independent, at least 2 forces or a non-smooth embedding of a parameter in the driving force are/is needed in order to change time dependence of MP.
A positive feedback is needed for MP to increase its absolute value with time (the same is for a price bubble). But a negative feedback is needed for MP to reduce its absolute value with time and hence for the mean-reverting. Change of FV (or assumption about FV number) opposite to MP time derivative can be a natural source for mean-reverting. In contrast with mean-reverting of natural and artificial systems probably many forces (co)-exist.
What are other possible fundamental forces for mean-reverting in stock market?