I think Fortran used ** because EBCDIC didn't have ^ or uparrow. ABC and Python followed Fortran rather than C on this point. units(1) supports both.
But IBM's BCD character sets, including the 48-character ones you allude to, didn't: https://en.wikipedia.org/wiki/BCD_(character_encoding)#Examp... (though Honeywell's did)
There are a lot of decisions in Fortran that stem from the absence of useful characters. .LT., .LE., .EQ., .NE., .GT., and .GE. is another.
I should have said "followed Fortran rather than BASIC".