This way, you know the host system is 100% going to represent what you want to say correctly, that the emoji will look right, and you don't need to worry the system will replace it, that it supports it etc..
And you can literally use 'any emoji' you want.
There is an infinite number of things we can put in emojis, why are we trying to standardize such things?
I wonder if someone has some insight as to the history around this?