Well, it almost certainly does not work "perfectly".
Given that the ICs are closely matched, the two MAX7219s are sharing the current from the Rset resistor, so it would be as if each has a resistor of twice that value and is driving half the nominal current to the displays. But it probably is not precisely shared, so one is likely to be driving more current than the other.
What is interesting is that given the non-linear response of the eye to light intensity, you will not have noticed, and if you correct the mistake, you will notice only a very modest apparent increase in brightness of the displays.