I stepped over them in the debugger watching R0. All functions returned zero, i.e. SSP_SUCCESS.
As stated above, I breakpointed the end result and looked at the various oscillator control registers and clock source registers. The appropriate oscillator stop bits were set. The system power mode was low-power mode. The subclock speed mode bit was set as stated above.
Also, the clockStop() calls for the main oscillator, MOCO, and LOCO are redundant. I make those same calls at the start of my application since I don't use those three clock sources.
The user manual tells me that subclock speed mode shuts down the HOCO. I pulled out the clockStop() and clockStart() calls on the HOCO after the last post. Measured current is still 1.7 uA as the user manual says it should be.
It was the accelerometer.
I think I've actually been getting the thing into subclock speed mode and into software standby mode for a day and a half. The 698 uA of accelerometer current made me think otherwise, however.