All _known_ serial killers isn't remotely the same thing as "all serial killers". Given that the United States is and has been roughly %5 of the world's population for the 20th century, the %75 factoid can't possibly reflect the reality. I'm not even sure where the %75 number comes from . . . whatever the case, it can't be right.
Have you ever heard of a serial killer from Burkina Faso, Peru, or Myanmar? From Zaire, Egypt or Mexico? From India or China? They surely exist/have existed, and there are a few cases that get reported -- but who's finding them? Which journalists are writing them up?
In order to be known as a serial murderer, two conditions are necessary: first the murderer must be able to carry out his murder and _not_ be identified and apprehended. Second, he must later be identified and apprehended, if not already dead. So you have a paradoxical situation-- a serial killer can operate freely when the ability to detect him is low, but we'll only know of him when we can actually detect him.
The US has a particularly active media and police interest in serial murder, producing a perception that dwarfs the reality, conflating murders with unknown perpetrators into serial murder.
Whatever the actual increase in the prevalence of serial murder in recent years, it is clear that fear associated with such crimes has grown. Prompted by exaggerated media reports (e.g., Darrach and Norris 1984), the American public has been scared into believing that there is an epidemic of serial murder in the United States, totaling as many as 5,000 victims annually.
{snip}
At some juncture, "unknown motive" was equated and confused with "no motive," leading to the erroneous inference that serial murder claims 5,000 victims per year (see Fox and Levin 1985; Jenkins 1988, 1994). Even when the flawed reasoning was uncovered, there remained a tendency to inflate uncritically the extent of the serial murder problem. When asked how many of the 5,000 homicides with unknown motives could be the work of serial killers, Justice Department sources speculated it to be two-thirds of the 5,000, or approximately 3,500 (Starr 1984).
Sources:
Multiple Homicide: Patterns of Serial and Mass Murder