Maybe, but all I can do is speculate based on the fact that the delay between when the teleport moves the player on client and when it is yanked back to that exact same location grows at a completely predictable linear rate.
I am having trouble imagining any other explanation than something is accumulating. Either time value is accumulating over time, or there is a delta between 2 times that is supposed to constant, but one is advancing a hair differently than the other and they grow more and more out of sync over time.
Another observation is that at larger degrees of rubber banding, I can teleport to a location and begin to walk away and stop before the rubber banding happens. I can just sit there as long as I want and I won't rubber band until I actually try to move again, and only then am I yanked back. Even after a long wait.
Another test I tried was to intentionally create a higher lag scenario. I logged into the server location on the far side of the world and with the highest ping, over 250 ms. With known lag what I observed was the lag caused problems with the beginning of the teleport, not the end. I would teleport after a small delay but I would not rubber band with the higher latency. Only after playing 10 or 15 minutes again did I see the rubber banding happen again at the end of the teleport, just like would happen with the 23 MS connection.