It depends as to the type of energy being radiated . All electrical devices create an electo-magnetic field when current flows through them. The CE testing is there to ensure that that radiation is kept to a minimum and devices are suitably sheilded against it. I believe that it is an offence to sell devices that do not meet this standard in europe. If it is radiated rf energy then the low voltage cable from the switched mode transformer to the cable modem can pick this up like a mini anttena. A ferite ring wound around the cable can help supres this.
Since the coax cable on the modem would filter the majority of rf interference the most likely source is reflected rf noise or pulses traveling through the mains system back to the equipment. I decent surge supresor with rf filtering could help stop this happening. After all everyone should have a surge supresor/filter as an absolute minimum, A UPS would be even better
It would be interesting to get a set of these dodgey lights and watch the mains supply on a fast sample scope for pulses or noise. The srarp increase and decrease of current that is drawn could well create pulses that cause the problem....
If the problem ocurs when the lights are some distance away then I would go with the noise/pulses down the mains supply theory.
HTH.