For a moment, let's assume that spell damage calculations in FF11 worked the same way in FF14. How M.Atk worked in FF11 was:
Base Damage (influenced from stats such as INT) x Multiplier (from stats like M.Atk)
Let us assume for a moment that a person with 100 INT (with no influences from M.Atk) does 100 damage with Fire, and M.Atk will change this damage result by a certain percentage. So let's say in addition to the earlier stats someone has an additional 200 M.Atk. This will change Fire's damage to 100 x 200% = 200 damage. Keep in mind this is all theoretical and simplified to make my point across, but FF11, in general, followed such a pattern of calculations.
Under such calculations, there will be profound differences in spell damage if one increases their M.Atk from 100 INT compared to that of someone increasing their M.Atk from 110 INT. To illustrate my point, let's use the earlier calculation methods by examining each case where one has different INT values but increases their M.Atk values by the same amount. Damages will be:
INT 100 x 200% = 200 Damage >>> INT 100 x 300 = 300 Damage
300 - 200 = 100. Therefore, an increase of 100 damage.
INT 110 x 200% = 220 Damage >>> INT 110 x 300 = 330 Damage
330 - 220 = 110. Therefore, an increase of 110 damage.
From this, we can come to the conclusion that if the game followed damage calculations through this kind of pattern, raising M.Atk will be the better choice as it offers the player a multiplier to damage as opposed to a straight increase. This is the reason why I took data for Case 2, because if it showed that M.Atk was a multiplier instead of a steady increase, the increase from M.Atk will be sharper for the person with a higher INT.
However, from the graph comparing the results from Case 1, it is shown that even when the M.Atk values vary, there is no increased change in damage. This might seem like a no-brainer to some because FF11 and FF14 are very different games, but because this issue has always been bothering me I'm glad it's finally put to rest.