While your answer is correct, Jim, it may over-complicate the problem. Most lasers operate at currents of at least 40mA, and as a result have very low intrinsic dynamic impedances. Taking this as an assumption, I matched my wideband 50-ohm RF signal to a Hitachi laser diode with a small 47-ohm SMD series resistor, and added a Picosecond Pulse Labs bias-T to inject the dc current. My network analyzer and TDR tests showed a reasonable 50-ohm termination was obtained. Overall light throughput response was flat, dropping by 1dB at 1200MHz. It should be noted I took special care to maintain effective double-terminated 50-ohm transmission lines for both the laser and PIN-diode receiver. It's possible the latter was responsible for the 1.2GHz rolloff. I noted with amusement that Hitachi's datasheet curves showed a peak at 600MHz, and dropped off rapidly thereafter. I surmised this was the response of their optical detector, or due to bad laser wiring.
Hint 1: All my open-wiring distances were kept to under 1mm. Hint 2: Learn about making high-performance bias-T networks. Pazzeo can read the recent discussion here, and photo postings on a.b.s.e.
Ahem. Pazzeo can "do the math" by replacing my 50-ohm resistor with a 75-ohm part. :>) But as to whether he can make good transmission lines and evaluate them, that may be another matter.