The article linked ... gives sufficient argument against the theory for anyone familiar with reality

I didn't find that so, myself. Yes, there were arguments, but there were a lot of assumptions behind the arguments that to me didn't stand up to too much scrutiny. But let's rephrase the argument a little. Hypothesis: Retaining the older staff has no impact on the employment opportunities of new entrants. Extremely unlikely, I would say, because it could only be true if opening that vacancy led to the loss of a vacancy elsewhere, such as in the restaurants mentioned. So in terms of employment I would say there must be a cost, even if it is not 1:1, which a naïve interpretation of the theory would suggest.

Then, in most countries in Europe - and maybe the US but I don't know enough about it - what you save on paying pensions you may well spend on Jobseeker's allowance, housing benefit, .... so when we think of employment we need to be thinking wider than business interests.

However, the discussion is about economics, not employment, so things are a little subtler, and I understand John's point that you cannot typically replace a long serving employee which knowledge of the business with an outsider who has no experience of the building. But it would be a foolish business - of which there are unfortunately plenty - who would try and do that. Instead, promote intermediates to draw on their experience. Of course, this requires businesses to think about succession planning and things like that, which too many see as overheads. But failure to do so by keeping older staff is short-termism at its worst: instead of a planned handover of skills from the older workers to less experienced staff, you keep Joe until he drops dead on the way to a meeting in a completely unplanned fashion, taking all his skills and experience with him and leaving no-one trained to take over.