The 1818 style turbos (or a TD05 which flows similar) seem to make best power on a 2.0 Subaru on Optimax at about 340 BHP and 1.25 bar on the more sensible dynos, and this is with tubular headers. When I've pushed them harder at least on the Subaru engines they start to lose power as you end up with really high exhaust manifold pressures, high EGTs, retarded timing and a wastegate blowing open (and it is usually not scratching where it is itching to use a stronger spring to keep it closed, think airflow not pressure). To then increase this to 400 BHP you'll need more airflow, (400/340*(1.25+1))-1 = 1.65 bar - and that is if you keep the same VE the same which is very unlikely with the same turbine wheel, and you'll have to run that at peak power RPM, and you'll have to be able to ward off detonation without excessive EGTs on your chosen fuel, and also that you can achieve similar charge temperature, or make even more boost to compensate... and you are in the spiral of asking ever too much from small turbos. The P20 will
help you a little bit, but you are trying to bridge a big gap, another 60+ BHP at this level is going to cost you big time in lag or it just won't happen if you stay too small.
The 1.8T probably is a bit more efficient so if might be compared to the EJ20 despite the capacity disadvantage, but I think you are aiming way too small to get an easy 400 BHP. Dig out some turbine wheel sizes and compressor maps and you'll see what I mean. Ignore the US dyno plots of course as they said for my terminal on the quarter I needed 440 WHP when I really had 350 WHP. I think you can virtually take their WHP figures as equivalent to our flywheel figures on Subarus!
Why not just leave it as it is?
Put it this way, I was following a 340 BHP Scooby (driven by a superior driver who also had better suspension) in the wet the other day, and he got away from me in 2nd and 3rd gear because he had traction and I didn't despite his approx 140 BHP and 0.5 litre deficit, and this is with four wheel drive on tyres that are decent in the wet. And then even with the capacity advantage, because the rotating assembly on my turbo is so much larger, it is silly speeds before the apparently large advantage is actually seen, feels far more dramatic, but the ground covering abiliity is not dramatically different.
You are, IMHO, in real danger of working harder and breaking stuff to actually go slower?
I know I'm always the voice of cynicism, but if you had a ride in the cars you would realise why, BTW you are welcome if you are ever up here to do so.