Hello Robin

I hope you will not mind if use your reference to Marshall's Very Long Run
as a pretext to raise some thoughts on what we might call the Very Very Long
Run.

During my attempts to make sense of early monetary history, I have chanced
to take an especial interest in the Afghan coinage of Alu-ud-Din Muhammed
Khwarezm Shah, c. 1200-20 AD, and that of the Chinese Emperor Wang Mang c.
7-23 AD

On the basis of the evidence we have, I would claim that Alu-ud-Din created
a proto-Hayekian monetary system, and Wang Mang a proto-Keynsian monetary
system.  

These putative facts remain little known.   I believe I was the first to
publish this claim about Alu-ud-Din, close on 20 years ago, and thus far
no-one has criticised my interpretation, indeed, no one seems to have
noticed it at all.  The claim concerning Wang Mang was in essence made by a
Chinese nationalist politician early in the 20th century, and seem to be
correct, despite numerous attempts by subsequent Western scholars to evade
the conclusion.

The odd fact about these two episodes, is that they both led to somewhat
similar outcomes.  Perhaps 50% of both populations were butchered by
barbarian hordes.

In the case of the Afghans, it appears that the central government so lost
support amongst the general population that it became impossible to organise
an effective defence against external barbarians, the army of Ghengis Khan.

In the case of the Chinese, the civil war between factions supporting fiat
currency and direct taxation, and those supporting indirect taxation and a
return to the gold standard, so wrecked the economy that anarchist peasant
armies, internal barbarians, with no policies whatever, rose up and killed
(and perhaps ate) a good proportion of the population, whilst many more just
starved to death.

I suppose that very very long run studies can only give us very very general
advice, but that advice would seem to be that moderation is a good plan, in
practice.

Apologies that I rather hijacked your own specific point  - this does seem
to me somewhat relevant to the general discussion 

Rob Tye