If you think the annual rate of catastrophic risk is X this century but only 0.5X next century, 0.25X the century afterwards etc, then you’d vastly underestimate expected value of the future if you use current x-risk levels to set the long-term discount rate.
Whether this is practically relevant or not is of course significantly more debatable.
If you think the annual rate of catastrophic risk is X this century but only 0.5X next century, 0.25X the century afterwards etc, then you’d vastly underestimate expected value of the future if you use current x-risk levels to set the long-term discount rate.
Whether this is practically relevant or not is of course significantly more debatable.
But the underestimation would be temporal.
I’m also not clear on whether this is practically relevant or not.