Ok, I have had some time to try a few more things.
I built a 50 ohm dummy load that can also be used to measure power. I measured 3.3 to 3.5 watts dead key depending on the channel, swinging to 8 watts or so with audio. So that seems reasonable.
SWR of 1.0 in to the dummy so I guess my meter is accurate, at least at a perfect match.
I tried playing with ground straps on the antenna mount and coax end but did not see any changes. I did notice that even with the electrical ground for the radio disconnected it all still works. Must be getting ground through the chassis or coax.
I also added a spring between my stud mount and the antenna. With this I measure a 1.8 SWR on 19 and 2.4 on the ends of the band.
With my other radio, powered through the lighter socket, I see a 1.4 SWR according to both the internal meter and my external meter. Also tested the internal meter with my dummy and read 1.00. Power is lower out of this radio, about 2.5W deadkey. Any thoughts on why this radio produces a lower SWR? Power output shouldn't effect SWR, correct? Should the radio chassis not be grounded?
Made a few contacts around town up to about 2 or 2.5 miles, which may be good enough.