I went back to the gospel (The manual for the X-11 FORTRAN program entitled THE X-11 VARIANT OF THE CENSUS METHOD II SEASONAL ADJUSTMENT PROGRAM), issued by the US Dept of Commerce back in 1967, written (I've just noticed) by Julius Shiskin, a remarkable econometrician. I examined the weights/algorithms I had been using in my seasonal adjustment program, and realised that when I wrote it 30+ years ago, I used several short cuts to get it up and running quickly. In the way these thing happen, I never got around to replacing the quick and dirty fixes I'd made with a more, shall we say, thorough working.
Anyway, this is what seasonally adjusted world EV/PHEV sales look like now (I haven't added the new data for March--I'll do that chart tomorrow. It's late, for me, and I'm off to bed shortly). As you can see, the values for January and February this year are somewhat higher than they were.
I just thought I'd let you know, as I said I would examine my program for flaws. Which took longer than expected because I had to work out anew what I'd done the first time, and I kept on thinking WTF? Anyway, all done.