I'd have to think that any water mill over a couple hundred years old would give a circa 2000 solar panel a run for its money; they weren't particularly expensive, and amortization adds up.
The amount of power your traditional water mill extracts would be exceeded by solar panels on its roof, which would be cheaper than the building under them, not even counting the cost of the water wheel and weir. The mill might last centuries, but only with continuous maintenance.
A weir way uphill with a penstock and Pelton wheel could do better than the old mill, but both depend on landform features with limited distribution. If you needed twice the power, you would be stuck. But you can put out more panels.
It's not clear to me why the reference points need to be a couple hundred years ago, and a couple decades ago. Was the intent to point out that the dollars are a one-off, where the kWh's keep on coming as long as the equipment works?
Even in my relatively wet and grey corner of the world, I'm confident that if I spent $X on PV vs hydro, the PV would produce more electricity and require less maintenance.
