Well I'm a bit out of touch now but I do know that each engine is tuned at manufacture, the nozzles (bits that direct the combustion gasses onto the turbine blades) are selected to give a compromise between torque and TOT margin. So you can have a good torque margin (gasses directed with little divergence) or a good TOT margins (gasses diverging more = cooler) but not both.
Historically with Makilas the TOT margin tended to get worse with age so they erred on the side of giving new engines a good TOT margin. But in fact the 2A1 seems to suffer more from worsening torque margins, so some time ago they altered the adjustment to favour torque margin. At least I think it was that way round!
Anyway if the torque margin is low as you describe I'd suspect the torque measuring system. Although it can be "calibrated" in the field in reality this seems fairly crude. It is certainly the case that you can move and engine from one aircraft to another and get a significantly different torque margins.
But an engine with negative torque margin is un-airworthy and should not be accepted by pilots or engineering. If it is on power by the hour it's easy, replace the engine.
However you mention that the torque margin has scatter with an average around zero, this is probably OK as the EPC can be subject to inaccuracies caused by heavy rain, high humidity or slight intake icing. One side tends to have slightly worse figures than the other due to installation issues (down/side wash from the rotor system) but I can't remember which side is worse (think it's LH side).
So I wouldn't worry about it too much if the average is zero. Make sure the engine wash procedure is done correctly with de-ionised water and that the pipes up to the spray rings aren't blocked.
Don't forget it's only an issue if you suffer an engine failure on the "better" engine at a critical stage of flight and as we know, that only ever happens in the simulator! By which I mean that if you ever have an accident, it won't be due to a torque margin of -1%.
Editing to say that air leaks from the bleed valve / heater takeoff area are a common cause of low torque margin but looking at the history of the EPCs usually shows a step change when the leak first occurred.
If the history shows that the engine has never been much good, and it's just dropped a % or so, I wouldn't worry. Worry about step changes.
I've done a lot of acceptance flights from the factory and some of the brand new aircraft had a torque margin of less than 1% but we had to accept them since it just has to be non-negative.