CGA on the original IBM PC produced sixteen colors, one bit each for red, green, blue and overall intensity modifier.
The preferred output device was the later-arriving 5153 color monitor, which accepted the RGBI signal in digital form, four bits (each on a separate wire, using TTL voltage levels) and performed its own digital-analog conversion.
This is a contrast with later VGA, where DAC was on the video card and the monitor accepted analog signals. Of course, the number of bits per pixel in VGA was much greater, so it's easy to see why VGA did DAC on the card.
Why did CGA leave DAC to the monitor? One possible answer would be that because there are only four bits per pixel, there is no particular disadvantage, so basically: why not.
But the other CGA output option was NTSC, and that involved doing DAC on the card after all. Admittedly not to the same analog format as the monitor ends up using, but it still seems intuitively likely that some of the circuitry could've been shared. Which in turn suggests there should be some offsetting positive advantage to leaving DAC in the monitor.
So why did CGA leave DAC to the monitor instead of putting it in the video card?