This observation comes from Tony Hoare’s paper “How Did Software Get So Reliable Without Proof?”. It draws on a broad pattern from commerce and industry: error rates tend to fall sharply once the cost of a mistake is borne by whoever caused it, rather than passed on to the customer.

The reasoning is about incentives. When those who make errors also pay for them, through rework, liability or lost reputation, they have a strong motive to prevent them in the first place. When the burden falls instead on the customer, that pressure is weakened and defects are more likely to persist.

In the context of the paper, Hoare applies this idea to software. He suggests that shifting responsibility for faults back towards their source is one of the practical mechanisms by which the industry improved reliability, complementing the more formal methods he had long advocated.