Here are the risks:
AFR - the ECU will ‘see’ the increased airflow, and will add fuel to maintain the target AFR. If you don’t exceed the limits of the stock fuel system, you might actually be OK on this front. With some years of STI, it’s common to reach the limits with a catless exhaust; I have no idea about the 08 or 09 WRX though. Lean AFRs tend to lead to knock, which tends to lead to broken pistons.
Timing - The ECU has tables that govern timing; these tables have RPM on one axis and load (meaning, grams of air ingested per intake stroke) on the other axis. These tables will reduce timing with increased load. But when you reach the edge of the table, the ECU uses whatever timing was specified at the edge of the table. With an 02 WRX, these tables go to 2.1 g/rev, and a catless exhaust will take you to 2.3. Dunno about the 08/09 WRX, but I wouldn’t want to bet my motor on it. Excessive spark advance tends to lead to knock, which again leads to broken pistons.
Boost - boost control requires a balance between the wastegate duty-cycle and the various restrictions in the intake and exhaust plumbing. Reducing backpressure to near zero, without retuning the WGDC tables, will result in higher boost. Excess boost tends to lead to one or both of the problems described above.
Your best-case scenario would probably just have you hitting the ECU’s fuel-cut over-boost protection every time you go full throttle. When you encounter that, it feels like you hit something in the road because you go from full power to zero very abruptly. It’s no fun. Not necessarily harmful, why bet your motor on it?
Your worst-case scenario would have you running too much spark advance and/or too little fuel, knocking a lot, and cracking ringlands.
Knock, in our motors, isn’t something you’re likely to notice while driving, so when someone says their motor is running fine after they’ve done something silly like this, you always have to imagine a “…so far” at the end of every claim they make about having a working motor.
Here’s a thread started by someone who installed a turbo-back exhaust without tuning:
Notice that IAM has dropped from its usual value of 16. The ECU doesn’t know that the car has been modded, so it assumes that all the detonation was caused by running crap fuel, so it reduces timing across the board.
Notice also that the ECU is experimenting with increased timing - see the positive numbers in the knock learning table? The ECU is hoping that there’s good fuel in the tank now. But this isn’t a fuel problem, it’s a tuning problem. The ECU will advance timing until detonation occurs again. It will then subtract timing, but only temporarily… it will continue trying to run stock timing, so it will continue detonating until the ECU is retuned for the car’s new exhaust.
I don’t mean to single out the person who posted that thread. Threads like that one crop up on LegacyGT.com and NASIOC.com and RomRaider.com every few months or so.