I am wondering to which extent versions of the DDM formula(https://en.wikipedia.org/wiki/Dividend_discount_model) is used by large banks and institutional investors etc, in some sense anyone who is not a retail investor.
Dividend discount model is flawed in the format it is given in.
It requires you to enter dividend (known), growth rate for dividends (well, you could estimate that for an individual stock but your estimate could be way off because this is predicting the future which is hard) and yield demand (this is very hard to give).
Specifically, the problem is with yield demand. By slightly adjusting the yield demand, you get very wildly different valuations. It is hard to justify what is the correct yield demand for a company. It will depend, among other things, on the riskiness of the company as an investment. Riskier companies require more yield. How much more, that's hard to specify.
So, the format
P = D/(r - g) is not really useful.
However, adjust the formula little and you will get
r = g + D/P
What does it tell you? It tells that "total expected return equates to dividend growth plus dividend yield".
In this format, the formula is very useful, because it tells you how much stocks yield. They yield dividend growth plus dividend yield.
In a large well-diversified portfolio, the dividend growth is nominal economic growth. It is useful to split it into two constituents: inflation and real economic growth. If doing that, the formula says "total expected return from a well-diversified stock portfolio equates to inflation plus real economic growth plus dividend yield".
That, to me, is a very strong and useful result. It can be used to estimate whether stock investments make sense, and if so, to what extent. The only difficulty of using that strong result is that you need to estimate what the cyclically averaged sustainable dividend yield is at a particular point of time.