The issue is not stability, but bass response. The capacitor drives the 259K grid resistor of the 2A3. I normally set the corner frequency pole at around 5Hz, which has worked well in many amps. Reducing the capacitance to 0.022uF would raise that point to 29Hz. (At the pole frequency the signal is 3dB down and 90 degrees out of phase with the midband.)
There are many arguments about the best approach. Some apply a phase shift rule and conclude the pole should be at 2Hz or lower; some say it should be higher to avoid driving the output transformer into saturation on music with heavy bass transients.
A point more rarely mentioned is that the recovery time from transient overload (grid current in the output tube) is proportional to the capacitance. A smaller capacitance recovers more rapidly. I have not seen any reliable data on the audibility of recovery time, though.
In short, it's worth trying. There's not much bass below 30Hz, after all. Remember, most teflon caps seem to need 100-500 hours of music to reach their best sound.
