10% average return on investment includes all slaves. Obviously some slaves were really bad investments. They got sick and died, for instance, and there goes your money. Or they ran away.
One factor that allowed slavery to thrive in the cotton south was the relatively even distribution of work over the course of a year. There was no incentive to hire workers just for the harvest, as you needed labor all year.
Slaveowners weren't stupid. They were well aware of the risk and rewards of the institution, and they chose to continue it. Slave prices generally increased from 1800 to 1860, although fluctuating with economic factors such as depression and cotton prices. Slave prices reached their peak in 1860, about $1500 for a prime field hand, at a time when the per capita GDP (according to one source, although it sounds low to me) was $137.
That buyers were willing to pay these prices is an excellent indication that they saw the purchase of a slave as a good investment.
If per-capita GDP is $137, how can a slave which cost $1,500 offer a 10% ROI?
That slavery continued in the South suggests that it was at least slightly better than the proposition of bringing in immigrant laborers, but it wouldn't have had to have been much better. The biggest disadvantage I could see to immigrant labor would be the difficulty of making a smooth transition, since nobody who had slaves would have wanted to take a major hit on their investment.
Market inertia can be strong; I'm curious to what extent slavery persisted because it was economically superior to immigrant labor, and to what extent it persisted because of inertia.