Note, in the figure the curves for above or below average SEB are computed at fixed (mean) IQ and vice versa. (So, e.g., one curve describes people with average IQ but above average SEB; another describes people with average SEB but above average IQ, etc.) It is only a coincidence (see last sentence in excerpt below) that the effect on age 19 income of being above or below average in SEB is approximately the same as for IQ. Were it not for this numerical coincidence the two high curves and the two low curves would not coincide at the age 19 intercept.
IIRC (from some earlier work), controlling for IQ (AFQT score) in this dataset (NLSY) eliminates almost all the earnings differential between blacks and whites.
A dynamic analysis of the effects of intelligence and socioeconomic background on job-market success
Yoav Ganzach, Faculty of Management, Tel Aviv University, Tel Aviv, Israel
Intelligence 39 (2011) 120–129
... The data were taken from the National Longitudinal Survey of Youth (NLSY), conducted with a probability sample of 12,686 Americans (with an oversampling of Afro-Americans, Hispanics and economically disadvantaged whites) born between 1957 and 1964.
... The measure of intelligence in the NLSY is derived from participants' test scores in the Armed Forces Qualifying Test (AFQT).
... four variables as indicators of SEB: education of the two parents, parent's family income, and occupational status of the parent holding the higher occupation.
... Fig. 2 provides a graphical representation of trajectories of participants high (one standard deviation above the mean) and low (one standard deviation below the mean) on intelligence and SEB, keeping the other time- invariant variables constant at their means. It is clear from the figure that while the trajectories of high and low intelligence diverge, the trajectories of high and low SEB do not. This suggests that intelligence, but not SEB, affected the slope of wage trajectories. It also suggests that SEB affected entry wages, and that its effect was similar to the effect of intelligence.
The effects of the two characteristics on the intercept, or entry level pay, are rather similar. On the basis of the multi-level parameters, one standard deviation increase in SEB [intelligence] leads to a 4.6% [5.4%] increase in entry pay.