Some stylistic reasons can be given.
(1) Lets admit that the verb is " pouvoir observer" . Putting the direct object complement before the the verb creates an equilibrium in the sentence, for the circumstantial complement ( after the verb) is very long. So the construction
Subject + Direct object ("le") + Verb + Circumstantial completement
allows to counterbalance the circumstantial complement.
(2) Second reason: if the writer had chosen this sentence construction :
"J’estime qu’on peut l'observer, au bout d’un temps plus ou moins long, dans les mariages d’amour"
the stress had been put on the circumstantial complement, I mean that the sentence could have been read as an answer to the question : " In what circumstances can this phenomenon be observed". And in that case, the answer could have benn interpreted as " it is not observed in all circumstances, but only sometimes, in certain special circumstances."
So the first construction tends to reduce the importance or the reality of the phenomenon. But this goes in a direction exactly opposite to the author's intentions.
What the writer aimed at was to emphasize the importance or reality of the phenomenon.
The question the writer wanted to answer was : " Is this phenomenon really observable ? ". And he wanted to stress the actual observability ( and therefore , reality) of the phenomenon.
Answer : " oui, on le peut observer" ( oui, il est effectivement observable)
With the second construction ( actually chosen), the circumstantial complement appears as a simple concession: the fact that the phenomenon is observable specially in these circumstances changes nothing to the fact that " on le peut observer".