It just wasn't anywhere near as bad as Harris claimed and if fitted with the big vertical stabilizers earlier operational losses might have been lowered sooner than waiting for the MK III.
Regarding the early Halifaxes, before the Mk.III, I'd be inclined to agree with Harris. We forget just how bad the early Halifax was because of its overall war record and how much of an improvement the Mk.III was over its predecessors - the Mk.III was a good aircraft and exhibited none of the dangerous characteristics and poor performance of the early examples; there was a war on and heavy bombers were required, so despite its many failings, the Hali was pressed on with. In the intervening years its many issues have been glossed over, but it was a basket case. Whilst undergoing trials at Boscombe Down, the service examples demonstrated a loss of speed of 20 mph over the prototypes because of turrets etc, maximum range during trials was decreased from 1,850 miles to 1,700 miles, rudder overbalance managed to destroy one aircraft, killing its crew - this issue alone took many hours and lives attempting to rectify over the subsequent couple of years. The scale of the effort undertaken to fix the type's persistent niggles took three years, in which time it consistently under performed, highlighted more so once the Lancaster began entering trial and service use, and in individual aircraft demonstrated dangerous handling, and shoddy workmanship, resulting in major performance deficiencies. This effort can be seen in the large number of types and sub-types of the Merlin engine examples that were built and the relatively small number of each that were built, after modifications were made on the production line after findings within the A&AEE.
It really was that bad. A good book to read detailing the issues surrounding the Merlin installation on the Halifax is
Rolls-Royce and the Halifax by David Birch.
Here's a quote from
The Secret Years, Flight Testing at Boscombe Down by Tim Mason;
"During 1942 and later the search continued for a solution to the problem of rudder overbalance thought in view of the Establishment's experience on the Mk.Is to be the cause of some unexplained service accidents and operational losses. The worst case considered was failure (and feathering) of the two port engines with the two starboard engines at take off power. The first modification, late in 1941, of reducing the rudder trimmer movement resulted in excessive foot loads until, on reducing speed to 160mph the rudders violently overbalanced to full travel and the resulting uncontrolled manoeuvre could only be overcome by throttling the starboard engines. 'Noseings', (bulbous nose to the rudders) improved matters such that control could be maintained down to a speed of 140mph with 10 deg bank without over balancing, but with very high foot loads.
Further changes, in mid 1942, to the balance tab and rudder tab settings reduced the foot loads to acceptable proportions, but reintroduced a mild tendency to rudder overbalance. It was decided to check the effectiveness on a representative aircraft. W7197 arrived in December 1942 from 102 Squadron; on the first handling test flight on 4 February 1943, the aircraft crashed fatally. The top half of one rudder had detached in flight - attributed to a violent overbalance leading to loss of control. Further investigation on HR679 (the first production aircraft incorporating the full range of aerodynamic improvements known as Series IA in service, but with rudders similar to W7917). Cautious reduction of speed during steady side slips caused no indication of overbalance until at 120 mph the rudders suddenly moved to full travel of their own volition. The pilot regained control at 150 mph after easing the control column forward; 4,000 ft had been lost in the spiral dive. On a second attempt control was regained by opening up the engines on the inside of the spiral. Later tests with restricted rudder movement reduced the speed at which less violent rudder overbalance occurred.
Cords on the trailing edge were tried but removed after it was found that they had no effect on overbalance and again made the rudders excessively heavy. HR727 had rudders with smaller balance areas which overbalanced at smaller angles; the modification was rejected. It was decided that larger fins were the only effective cure..."
The book goes on to describe the issues surrounding its loss of performance.