Is there any explanation for why in Figure 4, the aggregate line is so much higher than the two lines that are supposed to the same thing for the two groups tha

Is there any explanation for why in Figure 4, the aggregate line is so much higher than the two lines that are supposed to the same thing for the two groups that the population described by the aggregate line neatly partitions into? EDIT: I may have answered that to myself by drawing some income distribution curves. I guess it's a plausible outcome if the P90 for "all" is closer to the male P90 while the P10 is closer to the female P10 with the two curves being significantly different (which is increasingly the case as one goes to the past).