OK, this is embarassing.
I just checked my cranks in the context of the Shimano recall, and discovered that they are 170mm, rather than 172.5mm, which I thought they were and had been using in my power meter settings for the past 4 years. I use Assioma PM pedals.
Have adjusted the crank length now in the Assioma app, and in my Garmin.
What is the likely effect of the corrected crank length on measured power? (I won’t have the opportunity to do an FTP test over the weekend)
I think your power would’ve been out by 1.5% (that’s 2.5mm over 170mm). Nothing major to be concerned by, unless you had a perfectly round FTP 
f you set your cranks long and run on short cranks, you will overstate the power, and vice-versa. 172.5mm is 1.47% longer than 170mm.
torque = force x crank length, so power = rpm x force x crank length.
If the power meter thinks that the crank is 172.5, then it calculates power as X, say. If the actual crank length is 1.47% shorter than the power meter thinks, then the calculated power is 1.47% over the actual power. Which isn’t horrible.
Your FTP is actually 50+ watts higher than you thought it was! 
How coincidental! When I was testing something I moved from BLE to ANT+ and forgot to set the crank length too on Garmin, only noticed last night. Even the same crank length difference for me too.
Your FTP went down about 1.45%.
Luckily, you can buy your way out of this mess with taller socks!
When god closes a door, he opens a window.
Multiply your old FTP by 170/172.5 = 0.985 to get your real FTP.
For anyone that doesn’t know, the crank length that you set in the Assioma app only affects the Assioma app, nothing else. The pedals actually only transmit torque and whatever connects to them has to multiply by crank length. So you have to set the correct crank length in the head unit, TR app, or whatever you use to record or control using pedal power.