Hmm, bet the voltage drop across the LED is even higher. You could test with a JFET source follower, or JFET opamp follower, to see about what voltage it's at.
You would expect with, say, 8.2V in as shown, that "no" current flows, so R1 has 0V across it and the voltage between the diodes is undefined (anywhere between 0 and 8.2V). In reality, BAT45 is a very leaky diode, and will drop very little voltage relative to the LED, which might be a very good diode (i.e., low reverse leakage). Depending on ambient light level, that is.
Worth asking... did you measure this with real instruments (and if so, which), or in the simulation (that looks like... Multisim?)?
Tim